Datasets:

WINGNUS
/

ACL-OCL

	{
	"paper_id": "Y13-1000",
	"header": {
	"generated_with": "S2ORC 1.0.0",
	"date_generated": "2023-01-19T13:32:19.608574Z"
	},
	"title": "Supporting Organizations: Linguistic Society of Taiwan (LST) Association for Computational Linguistics and Chinese Language Processing (ACLCLP)",
	"authors": [
	{
	"first": "Alec",
	"middle": [],
	"last": "Marantz",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Huei-Ling",
	"middle": [],
	"last": "Lai",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Kawai",
	"middle": [],
	"last": "Chui",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Chao-Lin",
	"middle": [],
	"last": "Liu",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Shu-Chuan",
	"middle": [],
	"last": "Tseng",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Academia",
	"middle": [],
	"last": "Sinica",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Hee",
	"middle": [
	"-"
	],
	"last": "Chae",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Rahk",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Tsou",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Benjamin",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Min",
	"middle": [],
	"last": "Zhang",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Harada",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Yasunari",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Chiu-Yu",
	"middle": [],
	"last": "Tseng",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Rachel",
	"middle": [
	"("
	],
	"last": "Roxas",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "De",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Jong-Bok",
	"middle": [],
	"last": "Kim",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Bong",
	"middle": [
	"Yeung"
	],
	"last": "Lai",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Tom",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Jo",
	"middle": [
	"-"
	],
	"last": "Lin",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Wang",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Yanjun",
	"middle": [],
	"last": "Ma",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Baidu",
	"middle": [],
	"last": "Maekawa",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Ng",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Richard",
	"middle": [],
	"last": "Tsai",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Yuan",
	"middle": [],
	"last": "Tzong-Han",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Yuen-Hsien",
	"middle": [],
	"last": "Tseng",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Hui",
	"middle": [],
	"last": "Wang",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Cheng-Zen",
	"middle": [],
	"last": "Yang",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Jyi-Shane",
	"middle": [],
	"last": "Liu",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Siaw-Fong",
	"middle": [],
	"last": "Chung",
	"suffix": "",
	"affiliation": {},
	"email": ""
	},
	{
	"first": "Jason",
	"middle": [
	"S"
	],
	"last": "Chang",
	"suffix": "",
	"affiliation": {},
	"email": ""
	}
	],
	"year": "",
	"venue": null,
	"identifiers": {},
	"abstract": "Welcome Message from Conference Honorary Chair On behalf of ILAS, a co-host of this conference with National Chengchi University, I would like to welcome you all to the 27th Pacific Asia Conference on Language, Information and Computation (PACLIC27). The PACLIC conference has a long history,dating back to 1982 where the first conference of this series was organized with the original name \"Korea-Japan Joint Conference on Formal Linguistics\". It was the consensus of the organizer of the 1994 Joint Conference of the Asian Conference on Language, Information and Computation (ACLIC) and the Pacific Asia Conference on Formal and Computational Linguistics (PACFoCoL) that the two conferences would continue to be held jointly in the future as the Pacific Asia Conference on Language, Information and Computation, with the 1995 conference being numbered the 10th. Over the years the conference series has developed into one of the leading conferences in the Pacific-Asia region. Like the previous PACLIC conferences, PACLIC27 has received 123 submissions (workshop and main conference included) in the fields of theoretical and computational linguistics, and participants coming from 27 countries. The long tradition of the conference has been the keynote and invited speakers, and this year is no exception. The 5 eminent scholars who kindly agreed to deliver keynote speeches for this year are",
	"pdf_parse": {
	"paper_id": "Y13-1000",
	"_pdf_hash": "",
	"abstract": [
	{
	"text": "Welcome Message from Conference Honorary Chair On behalf of ILAS, a co-host of this conference with National Chengchi University, I would like to welcome you all to the 27th Pacific Asia Conference on Language, Information and Computation (PACLIC27). The PACLIC conference has a long history,dating back to 1982 where the first conference of this series was organized with the original name \"Korea-Japan Joint Conference on Formal Linguistics\". It was the consensus of the organizer of the 1994 Joint Conference of the Asian Conference on Language, Information and Computation (ACLIC) and the Pacific Asia Conference on Formal and Computational Linguistics (PACFoCoL) that the two conferences would continue to be held jointly in the future as the Pacific Asia Conference on Language, Information and Computation, with the 1995 conference being numbered the 10th. Over the years the conference series has developed into one of the leading conferences in the Pacific-Asia region. Like the previous PACLIC conferences, PACLIC27 has received 123 submissions (workshop and main conference included) in the fields of theoretical and computational linguistics, and participants coming from 27 countries. The long tradition of the conference has been the keynote and invited speakers, and this year is no exception. The 5 eminent scholars who kindly agreed to deliver keynote speeches for this year are",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	}
	],
	"body_text": [
	{
	"text": "Main",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Table of Contents",
	"sec_num": null
	},
	{
	"text": "Sparseness Problems in Word Sense Disambiguation: A Case for Korean and Its Applications to English and Chinese........................................................... 17 Aesun Yoon, Minho Kim, ",
	"cite_spans": [
	{
	"start": 92,
	"end": 185,
	"text": "English and Chinese........................................................... 17 Aesun Yoon,",
	"ref_id": null
	},
	{
	"start": 186,
	"end": 196,
	"text": "Minho Kim,",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Using Hierarchically Structured Lexicon as Key Clues Solving Data",
	"sec_num": "2."
	},
	{
	"text": "3. Head-Internal Relatives in Japanese as Rich Tohru Seraku",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "One-Soon Her and Yun-Ru Chen",
	"sec_num": null
	},
	{
	"text": "Oral Presentation Session 1B: Language Generation ",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "One-Soon Her and Yun-Ru Chen",
	"sec_num": null
	},
	{
	"text": "3. ChinGram: A TRALE Implementation of an HPSG Fragment of Mandarin",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "R. Zuber",
	"sec_num": null
	},
	{
	"text": "Chinese ...........................................................................................................",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "R. Zuber",
	"sec_num": null
	},
	{
	"text": "Oral Presentation Session 5B: Machine Translation Abstract Pinker's influential presentation of the distinction between the combinatoric units of language (the \"words\") and the mechanisms that organize the units into linguistic constituents (the \"rules\") rested on a strong, but ultimately incorrect, theory about the connection between a speaker's internalized grammar and his/her use of language: that what is linguistically complex, and thus constructed by the grammar, is not memorized; thus experience with complex constituents (as measured in corpus frequency, for example) would have no effect on processing such complex constituents. I argue that recent results within linguistics and within psycho-and neuro-linguistics show instead that memory and frequency effects are irrelevant to the linguistic analysis of language but always influence processing, across simple and complex constituents. Phrases and words can be shown always to decompose down to the level of morphemes both in representations and in processing, and, contrary to Pinker's claim, the \"memorized\" status of a complex structure holds no import for its linguistic analysis. On the other hand, speakers' experience with language is always reflected in their use of language, so frequency effects are always relevant to processing, even for completely regular combinations of words and morphemes. I will present neurolinguistic evidence for full decomposition of irregular forms (such as English irregular verbs), as well as evidence for frequency effects for regular combinations of morphemes and words. expressions. Many such expressions can be represented as rules or templates. These templates are matched by computer to identify those linguistic objects in text. However, in real world, there always seem to be many exceptions or variations not covered by rules or templates. A typical approach to cope with this situation is either to produce more templates or to relax the constraints of the templates (e.g., by inserting options or wild cards). But the former could create many similar case-by-case templates with no end in sight; and the latter could lead to lots of false positives, namely, matched but undesired linguistic expressions. Thus, the flexibility of rule matching has troubled the natural language processing (NLP) as well as the artificial intelligence (AI) community for years so as to make people believe that rule-based approach is not suitable for NLP or AI in general. On the other hand, fine-grained linguistic knowledge cannot be easily captured by current machine learning models, which resulted in mediocre recognition accuracy. Therefore, how to make the best out of rule-based and statistical approaches has been a very challenging task in natural language processing. This paper describes a partial matching scheme that enables a single template to match a lot of semantically similar expressions with high accuracy, which we refer to as the Principle-Based Approach (PBA).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "In PBA, we use a collection of frames to represent linguistic concepts or rules.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "Each frame is a collection of slots (also called components) with relations specified among them. A slot can be a word, phrase, semantic category, or another frame concept. One can specify position relations, collocation relations, and agreement relations and others among its slots. Unlike normal templates that involve mostly left-right relations among its components in a sentence, relations within frames can be multi-dimensional. For example, one slot could be a variable indicating the topic which other slots belong to.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "To illustrate our partial matching scheme, consider a simple frame concept involving 5 components such that their relations in a sentence are arranged as 1, 2, 3, 4, 5 from left to right. Suppose in a sentence we can identify components 2, 3, and 5 in that order. So 1 and 4 are missing (deletion), and there maybe words inserting between 2 and 3 (insertion), and also between 3 and 5. Furthermore, a match for slot 5 could be on word-sense rather than on the word themselves (substitution). Our partial matching scheme allows for insertion, deletion and substitution. An insertion is given a positive score if it tends to collocate with its left or right matched components in general (otherwise, negative). A deletion can be harmless if slots 2, 3, and 5 contain a key combination for the frame. Note that many such key combinations can be pre-specified as indices of the frame. Collocation and bigram statistics can be incorporated in such score estimation. A substitution is given a lower score depending on their closeness in a semantic tree. After all these scores are determined, we can use an alignment algorithm to measure the fitness score and to decide how well the frame matches with the sentence.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "PBA is inspired by the fact that when one studies a foreign language, he or she is usually presented with a collection of rules. These rules and their possible extensions and variations are practiced over and over again in real life to be mastered by the learner. PBA is flexible in that, it tends to relieve the burden of having to match with something \"exactly\" as specified and fine-grained linguistic knowledge can be more easily adopted to help estimate the scores of insertion, deletion and substitution in a PBA frame match.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "We believe PBA can model more linguistic phenomena than current machine learning models, and is more suitable for NLP and AI in general. More details and examples of PBA will be covered in the talk.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Stefan M\u00fcller and Janna Lipenkova",
	"sec_num": null
	},
	{
	"text": "Extracting \"Criterial Features\" for the CEFR Levels Using Corpora of EFL Learners' Written Essays Yukio Tono Tokyo University of Foreign Studies y.tono@tufs.ac.jp",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "PACLIC-27",
	"sec_num": null
	},
	{
	"text": "In this talk, I will report on the on-going project on systematic extraction of criterial features from multiple source corpora based on the Common European Framework of Reference for Languages (CEFR). First, a brief description of the CEFR itself, the project and the design of several different corpora newly compiled for the project will be given, followed by methodological issues regarding how to extract criterial features from CEFR-based corpora using machine learning techniques.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "The project aims to support the implementation of the CEFR-J, an adaptation of the CEFR into English language teaching in Japan (Tono & Negishi 2012) . After the release of Version 1 of the CEFR-J in March, 2012, we launched a new government-funded project called the \"CEFR-J Reference Level Description (CEFR-J RLD)\" Project. RLD is a term used for the CEFR to prepare an inventory of language (lexis and grammar) for each individual language for the purpose of level specification. students essay data with approximately 5,000 samples aligned with correction data.",
	"cite_spans": [
	{
	"start": 128,
	"end": 149,
	"text": "(Tono & Negishi 2012)",
	"ref_id": "BIBREF33"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "The other is the data collected by Ministry of Education (MEXT), in which more than 2,000 students were randomly selected from all over Japan. They were given written and oral proficiency exams in English. This data shows the average performance of EFL learners in Japan, after the three year instructions in secondary school.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "Finally, a corpus of classroom interaction between teachers and students has been added to the resource. This is an on-going project and the size is relatively small, but I hope that it will shed light on the understanding of what is happening in the classroom.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "Our aim is to identify criterial features by looking at input and output corpora across CEFR levels. The language presented in the input corpora may not be produced in the output corpora. By examining both input and output, descriptions of criterial features will become more systematic. The interaction corpus also helps better understand the learning/acquisition process in the classroom. Input from textbooks as well as input and interactions in the actual classroom will play an important role in learning a target language. The major goal is to find out criterial features for the levels specified in the CEFR-J and complete the inventory of grammar and vocabulary for teaching and assessment, with a special reference to teaching and learning contexts in Japan.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "In the past few years, various linguistic criteria have been proposed as \"criterial\", but they need to be validated against a particular learner group like Japanese EFL learners because the data used in Europe are very different from our learner group.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "Also each proposed criterial feature should be evaluated and weighed in terms of PACLIC-27 usefulness as CEFR-level \"classifiers\". Then a bundle of criterial features have to be tested and validated to find out which combinations of criterial features work best to predict the CEFR-levels. In a way, for assessment purposes, it is sufficient to identify the most salient criterial feature that can distinguish all the levels clearly. For teaching purposes, however, all the learning items need to be somehow evaluated against their 'criteriality.'",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "There are various ways of extracting criterial features from the data. Machine learning techniques such as random forest seem to be very promising for this purpose.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "For instance, random forest is very useful in that it gives estimates of what variables are important in the classification. Table 2 shows the results of variable importance measure by Gini impurity criterion. Basically, the higher the score is, the more important the variable is. By using this kind of information, one can profile which linguistic feature will be most effective in classifying texts into CEFR levels. The major aim of the project is to decide on which machine learning algorithms to take, and evaluate a range of criterial features for its effectiveness as assessment and teaching points. ",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 125,
	"end": 132,
	"text": "Table 2",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "The CEFR-J and Reference Level Descriptions",
	"sec_num": null
	},
	{
	"text": "By its very nature, corpus linguistics is a discipline not just concerned with, but ultimately based on, the distributions and frequencies of linguistic forms in and across corpora. This undisputed fact notwithstanding, for many years, corpus linguistics has been dominated by work that was limited in both computational and statistical ways.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "As for the former, a lot of work is based on a small number of ready-made proprietary software packages that provide some major functions but can of course not provide the functionality that, for instance, programming languages provide. As for the latter, a lot of work is very unstatistical in nature by relying on little more than observed frequencies or percentages/conditional probabilities of linguistic elements.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "However, over the last 10 years or so, this picture has changed and corpus linguistics has evolved considerably to a state where more diverse descriptive statistics and association measures as well as multifactorial regression modeling, other statistical classification techniques, and multivariate exploratory statistics have become quite common. In this talk, I will survey a variety of recent studies that showcase this new-developed methodological variety in both synchronic and diachronic corpus linguistics; examples will include applications of generalized linear (mixed-effects) models, different types of cluster-analytic algorithms, principal components analysis and other dimension-reduction tools, and others.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "Beijing, China jtsujii@microsoft.com",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Linking Text with Data and Knowledge Bases Junichi Tsujii Microsoft Research Asia",
	"sec_num": null
	},
	{
	"text": "In the last two decades, we have witnessed the rapid development of techniques in statistical modeling of language, which exploit large collections of text to reveal statistical regularities in language uses. However, the statistics-based approach to language,which tends to ignore or deemphasize structural issues of language, has shown its own limitations. The approach in its strictest form, for example, fails to With this broad perspective in mind, I will address the following research topics which I have been involved in:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "(1) Parsing and Semantics: While the performance of a syntactic parser has been improved substantially of late, it still fails to treat semantically crucial constructions. In order to resolve the difficulties which remain in parsing, we have to treat semantics of language more systematically than the current state of the arts parsers do. I would argue that we cannot resolve the difficulties without referring to proper theories of syntax.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "(2) Entity linking: Disambiguation in entity-linking has been carried out by using characteristics specific to individual entities. However, in order to treat long-tail problems in entity-linking, not only properties of individual entities but also classes of entities and their properties in knowledge bases have to be exploited.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "The results of our recent experiments will be presented, in order to illustrate how structures in knowledge bases can be used for interpretation of expressions in language.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "(3) Relation linking: The same relation in the knowledge domain can be expressed by diverse surface expressions in language. To gather surface relation expressions for a given set of relations in the knowledge domain is a crucial step of linking text with knowledge. Some of our recent studies in relation extraction will be presented as the next step of linking text with knowledge bases. derived from the British National Corpus.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "The n-grams resulting from the queries can then be linked to examples extracted from the New York Times Corpus in order to provide full sentential context for more effective learning. In some situations, users might need to search for words in a specific syntactic relation (i. e., collocates). Let's consider the example \"absorb $N\" that queries all the objects of the verb absorb. In this case, grouping the words that belong to similar domains together offers a better overview of the usage of the verb than a list of objects ordered by frequency. For example the verb absorb takes clusters of objects related to the topic liquid/energy, but also to the topics money, knowledge or population.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "This tendency of predicates to prefer certain classes is defined by Wilks (1978) ",
	"cite_spans": [
	{
	"start": 68,
	"end": 80,
	"text": "Wilks (1978)",
	"ref_id": "BIBREF42"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "As we have well known that it is always a basic requirement for statistical machine translation (SMT) to maintain semantic equivalence between a source sentence and its translation. However, nearly all of the existing translation models do not deal with the semantic structure between two languages at all. In this talk, I will present a novel translation method based on semantically-motivated framework, using predicate-argument structure (PAS). Generally, PAS depicts the semantic relation between a predicate and its associated arguments, and it always indicates the semantic frame and skeleton structure of a sentence. Thus, we believe the PAS would be much beneficial for machine translation in grasping the semantics of sentences. Furthermore, after analysis of the weakness of PAS representation during translation, I will propose a concept of syntax-complemented PAS (SC-PAS). It effectively overcomes the drawback of the prevalent gaps in PAS and provides more useful knowledge for SMT. ",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	},
	{
	"text": "The verbs meaning 'give' across languages are known to be among the most highly grammaticalized verbs, which exhibit a high degree of polyfunctionality. This paper aims to (i) present commonalities and differences in the grammaticalization of the verbs for 'give' in Thai and Mandarin Chinese, namely, h\u00e2y in Thai and g\u011bi in Mandarin Chinese, and (ii) investigate how different constituent orders of the head vis-\u00e0-vis the modifier and complement in Thai and Mandarin Chinese bear on patterns of grammaticalization of the two verbs. It is found that the functions that h\u00e2y in Thai and g\u011bi in Mandarin Chinese share in common are (1) the ditransitive verb use, (2) the dative-marking use, (3) the benefactivemarking use, and (4) the causative-marking use. As for different functions of h\u00e2y and g\u011bi, h\u00e2y exhibits the clause connective use, which is lacking in g\u011bi, whereas g\u011bi exhibit the passive-marking use, which is lacking in h\u00e2y. It is argued that the head-modifier order in Thai seems to be compatible with postverbal grammaticalized morphemes whereas the modifier-head order in Mandarin Chinese seems to be compatible with preverbal grammaticalized ones.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstr act",
	"sec_num": null
	},
	{
	"text": "It is generally known that Thai and Mandarin Chinese are typologically similar in many respects. They are isolating, topic-prominent, serializing, have the SVO basic word order and rich with grammaticalized morphemes. However, there is one important difference between them, i.e. difference in constituent order. Mandarin",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Intr oduction",
	"sec_num": "1"
	},
	{
	"text": "Chinese has the modifier-head order whereas Thai has the head-modifier one. This paper investigates how the difference in constituent order in Thai and Mandarin Chinese bears on patterns of grammaticalization of serial verbs in the two languages. The serial verbs for 'give' in Thai and Mandarin Chinese, i.e. h\u00e2y and g\u011bi, are used as a case study. The verbs meaning 'give' across languages are known to be among the most highly grammaticalized verbs, which exhibit a high degree of polyfunctionality. The analysis in this paper is based on the findings of a synchronic contrastive study of h\u00e2y and g\u011bi presented in Thepkanjana and Uehara (2008) .",
	"cite_spans": [
	{
	"start": 616,
	"end": 645,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Intr oduction",
	"sec_num": "1"
	},
	{
	"text": "2 Commonalities an d differ ences Thepkanjana and Uehara (2008) make a synchronic contrastive study of the polysemous morphemes h\u00e2y and g\u011bi in Thai and Mandarin Chinese. It is found in Thepkanjana and Uehara (2008) that h\u00e2y and g\u011bi share four main uses, namely, the ditransitive (main) verb use, the dative-marking use, the benefactive-marking use and the causative-marking use. As for differences between h\u00e2y and g\u011bi, one important use that is missing in h\u00e2y is the passive-marking use whereas one that is missing in g\u011bi is the clause connective function. The commonalities between the two verbs are discussed in section 2.1 and the differences in section 2.2. The examples provided are drawn from Thepkanjana and Uehara (2008) . Notice that the dative h\u00e2y in Thai occurs postverbally whereas the dative g\u011bi occurs both preverbally and postverbally. Newman (1993b) argues that an act of giving naturally results in some kind of benefit to the recipient. Even a non-giving action, such as driving, speaking and cleaning can also be done for the benefit of someone. The person who benefits from the agent's action is usually called a beneficiary. Therefore, it is natural that h\u00e2y and g\u011bi can also function as benefactive markers. The notion of benefactive is more complicated than generally assumed. Three types of benefactive are postulated in this paper as below.",
	"cite_spans": [
	{
	"start": 34,
	"end": 63,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 185,
	"end": 214,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 699,
	"end": 728,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 851,
	"end": 865,
	"text": "Newman (1993b)",
	"ref_id": "BIBREF47"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Intr oduction",
	"sec_num": "1"
	},
	{
	"text": "(a) Recipient benefactive: The beneficiary gains a benefit by virtue of being a recipient of a concrete entity, for example, John bought a sweater for Mar y. (b) Benefit benefactive: The beneficiary gains a more or less abstract benefit from somebody's action, for example, John sang a song for Mar y. (c) Behalf benefactive: The beneficiary gains a benefit from somebody who performs an action on his/her behalf, for example, John drove a car for Mar y because she was drunk.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "It is found that the Thai h\u00e2y can be used to mark the three types of benefactive as shown below. The structural schemas of the benefactive h\u00e2y and g\u011bi are given below.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "Benefactive-marking use Thai:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "EQUATION",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [
	{
	"start": 0,
	"end": 8,
	"text": "EQUATION",
	"ref_id": "EQREF",
	"raw_str": "[V + (NP1) + h\u00e2y + NP2]",
	"eq_num": "(beneficiary)"
	}
	],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "Mandarin Chinese:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "EQUATION",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [
	{
	"start": 0,
	"end": 8,
	"text": "EQUATION",
	"ref_id": "EQREF",
	"raw_str": "[g\u011bi + NP1 + V+ (NP2)]",
	"eq_num": "(beneficiary)"
	}
	],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "Notice that the benefactive h\u00e2y and g\u011bi occur in different positions. The former occurs postverbally, i.e. after the main verb, whereas the latter occurs preverbally, i.e. before the main verb.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "The third common use of h\u00e2y and g\u011bi is the causative use. The causative constructions with the causative-marking h\u00e2y and g\u011bi in Thai and Mandarin Chinese have the same syntactic schema as below. The NP1 in the schema above is the causer whereas the NP2 is the causee. The causer is typically human whereas the causee is typically animate. The causative verbs h\u00e2y and g\u011bi express an indirect causation in which the causer intentionally causes an event to take place by doing something to prompt the causer to act or by not doing something which prevents that event to take place. The causee is the person who directly causes the event to take place. Notice that the causative g\u011bi occurs in the same position as the benefactive g\u011bi in Mandarin Chinese, which results in ambiguity between the causative and benefactive readings in some cases as shown in (15), which is taken from Newman (1996:20) . 15 'S/he built a house for me.' (preferred) 'S/he had me build a house.' (awkward)",
	"cite_spans": [
	{
	"start": 877,
	"end": 893,
	"text": "Newman (1996:20)",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "Yap and Iwasaki (1998) note that Mandarin Chinese prefers the causative verbs r\u00e0ng and ji\u00e0o to the verb g\u011bi in expressing indirect causation as in (17).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "(17) t\u0101 *g\u011bi/r\u00e0ng/ji\u00e0o h\u00e1izi shu\u00ec-ji\u00e0o s/he CAUSE child sleep 'She let the child sleep.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "The use of r\u00e0ng and ji\u00e0o rather than g\u011bi to express causation helps prevent the ambiguity between the causative and benefactive readings that can arise if g\u011bi is used as the causative verb, which occurs in the same position as the benefactive g\u011bi. It is therefore not surprising that the use of the causative g\u011bi in Mandarin Chinese is much more restricted than the use of the causative h\u00e2y in Thai because the latter does not create ambiguity as the former.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Commonalities between h\u00e2y an d g\u011bi",
	"sec_num": "2.1"
	},
	{
	"text": "H\u00e2y and g\u011bi are different in two ways. There is one important use of h\u00e2y which is missing in g\u011bi, namely, clause connective use, and one important use of g\u011bi which is missing in h\u00e2y, namely passive-marking function. The clause connective use, which is missing in g\u011bi is discussed first.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Differ ences between h\u00e2y an d g\u011bi",
	"sec_num": "2.2"
	},
	{
	"text": "The connective h\u00e2y in Thai takes place in complex constructions in which h\u00e2y functions as a subordinator which links two predicates or two clauses. The first clause in the complex construction is the matrix clause and the other is the subordinate one. The complex constructions in which h\u00e2y functions as the subordinator can be classified into three types, namely, a purposive construction, a jussive construction and a complementation construction. The purposive construction is a complex construction in which the subordinate clause functions as a purpose of the performance of an action denoted by the matrix clause. The jussive construction expresses a command, request or demand made by one participant towards another in order for the latter to perform an action (Van Valin and LaPolla, 1997) . The complementation construction is a complex construction in which the subordinate clause functions as a complement of the desiderative predicate of the matrix clause. The structural schema of the connective h\u00e2y and some examples of the three types of complex constructions containing h\u00e2y are given below. phaefl aet study medicine 'Somsak wanted his child to study medicine.' Thepkanjana and Uehara (2008) argue that each type of complex construction results from a reanalysis of h\u00e2y from the causative verb to the subordinator. In the reanalysis process, the causative h\u00e2y is semantically bleached out and loses its verbal properties to varying degrees in the three types of complex construction. In other words, h\u00e2y in the three types of complex construction has different degrees of function word properties. It is argued in Thepkanjana and Uehara (2008) that the connective h\u00e2y in the complex constructions is derived, extended or grammaticalized from the causative h\u00e2y. The h\u00e2y's in all of these cases are followed by a clause or a predicate. The causative h\u00e2y functions as the main verb in the causative construction whereas the connective h\u00e2y is preceded by a main verb and followed by a clause or a predicate. It is found that there is an intention that an event take place in the subject of the matrix clause in all of the three types of complex construction and in the subject of the causative h\u00e2y. It is argued in Thepkanjana and Uehara (2008) that the notion of indirect causation has the highest degree of saliency in the causative h\u00e2y but has decreasing degrees of saliency in the purposive, jussive and complementation constructions.",
	"cite_spans": [
	{
	"start": 769,
	"end": 798,
	"text": "(Van Valin and LaPolla, 1997)",
	"ref_id": "BIBREF51"
	},
	{
	"start": 1179,
	"end": 1208,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 1631,
	"end": 1660,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 2228,
	"end": 2257,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Differ ences between h\u00e2y an d g\u011bi",
	"sec_num": "2.2"
	},
	{
	"text": "On the other hand, one important use of g\u011bi which is missing in h\u00e2y is the passive-marking function. The passive-marking function is alternatively called the agentive-marking function. The structural schema of the passivemarking g\u011bi and some examples are given below. According to Xu (1994) , the passive g\u011bi is used in colloquial speech whereas the other passive marker, b\u00e8i, is used in formal speech. In addition, a verb which co-occurs with the passive g\u011bi must be marked by the aspect marker le, otherwise the sentence with g\u011bi will not be interpreted as a passive sentence. Many works, such as Newman 1993a, b), Xu (1994) , Iwasaki (1998, 2003) argue correspondingly that the passive g\u011bi is directly derived from the causative g\u011bi via the reflexive context. An important question is why the development from a causative use into a passive one does not take place in Thai. Yap and Iwasaki (1998) found out that h\u00e2y in Thai takes only a volitional causer. Yap and Iwasaki (2003) argue that only nonvolitionality on the part of the causer can allow a passive interpretation to emerge. Therefore, the high degree of volitionality of the causer prevents h\u00e2y from developing into a passive marker in Thai.",
	"cite_spans": [
	{
	"start": 281,
	"end": 290,
	"text": "Xu (1994)",
	"ref_id": "BIBREF43"
	},
	{
	"start": 617,
	"end": 626,
	"text": "Xu (1994)",
	"ref_id": "BIBREF43"
	},
	{
	"start": 629,
	"end": 649,
	"text": "Iwasaki (1998, 2003)",
	"ref_id": null
	},
	{
	"start": 877,
	"end": 899,
	"text": "Yap and Iwasaki (1998)",
	"ref_id": "BIBREF44"
	},
	{
	"start": 959,
	"end": 981,
	"text": "Yap and Iwasaki (2003)",
	"ref_id": "BIBREF45"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Clause connective use",
	"sec_num": null
	},
	{
	"text": "In summary, h\u00e2y in Thai occurs in four constructions, namely, the ditransitive construction, the prepositional phrase, the causative construction and the complex construction. H\u00e2y functions as the ditransitive main verb, dative and benefactive markers, causative verb and clause connector or subordinator, respectively. Each of the four constructions has its own structural schema as below. The syntactic category of h\u00e2y in each construction and function is specified under each structural schema in the rightmost column. On the other hand, g\u011bi in Mandarin Chinese appears in four constructions, namely, the ditransitive construction, the prepositional phrase, the causative construction and the passive construction. G\u011bi functions as the ditransitive main verb, dative and benefactive markers, causative verb and passive marker, respectively. The constructions in which g\u011bi appears, the functions and the structural schemas of all constructions containing g\u011bi appear in Table 2 Some observations can be made regarding the functions, the structural schemas and the productivity of h\u00e2y and g\u011bi in the functions specified in the tables above as follows.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 971,
	"end": 978,
	"text": "Table 2",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "Summar y",
	"sec_num": "2.3"
	},
	{
	"text": "(a) The clause connector use is possible for h\u00e2y in Thai but is lacking for g\u011bi in Mandarin Chinese. (b) The passive-marking use in possible for g\u011bi in Mandarin Chinese but is lacking for h\u00e2y in Thai. (c) The g\u011bi-marked dative PP in Mandarin",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Summar y",
	"sec_num": "2.3"
	},
	{
	"text": "Chinese can occur both before and after the main VP whereas the h\u00e2y-marked dative PP can occur only after the main VP. That means there are two structural schemas of the dative g\u011bi whereas there is only one of the dative h\u00e2y. In section 3, we will argue for the relationship between constituent orders in Thai and Mandarin Chinese on the one hand and patterns of grammaticalization of h\u00e2y and g\u011bi on the other.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Summar y",
	"sec_num": "2.3"
	},
	{
	"text": "In this section, we will point out how constituent orders in Thai and Chinese bear on patterns of grammaticalization of h\u00e2y and g\u011bi in both languages. The constituent orders to be discussed in this section are those of a head vis-\u00e0-vis a modifier and those of a head vis-\u00e0-vis a complement. A complement is a syntactic category that is selected or subcategorized for by the head of a phrase. A complement is therefore semantically necessary for the head to become semantically complete. Some examples of complements are below.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "(26) I cut a tree. (27) She put a book on the table.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "In 26and (27), the direct object nominals a tree and a book function as complements of the verbs cut and put respectively. In addition, the prepositional phrase on the table also functions as another complement of the verb put in (27) because the verb put is semantically incomplete without it. On the other hand, a modifier is an expression which limits or qualifies the meaning of a word, a phrase or a sentence. It is less semantically crucial to the meaning of a head than a complement. In other words, a modifier is more semantically peripheral than a complement. The underlined parts in (28) and (30) illustrate the modifiers in the sentences. In (28), very modifies tall. In (29) and (30), the phrases in the living room and after dinner modify the predicates in the clauses. The three sentences above are semantically complete without the modifiers. However, Langacker (1987) acknowledges that the demarcation between modification and complementation is sometimes hard to draw because the difference between them is a matter of degree.",
	"cite_spans": [
	{
	"start": 867,
	"end": 883,
	"text": "Langacker (1987)",
	"ref_id": "BIBREF53"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "It is generally known that the constituent orders in Thai and Mandarin Chinese are different in that Thai has the head-modifier constituent order whereas Mandarin Chinese has the modifier-head one. The difference in constituent order in the two languages is illustrated below. The adverbial modifiers in the examples are underlined. However, in case of the head and complement, the constituent orders in Thai and Mandarin Chinese are identical, that is, head-complement order. Therefore, in Mandarin Chinese, the modifier appears before the head whereas the complement appears after the head. On the other hand, in Thai, both the modifier and the complement appear after the head. In this section, we will point out that the constituent orders of the head and modifier and of the head and complement in Thai and Mandarin Chinese have some effects on patterns of grammaticalization of h\u00e2y in Thai and g\u011bi in Mandarin Chinese. To be specific, we will provide answers to the following questions in terms of different constituent orders in Thai and Mandarin Chinese. occurs only preverbally. We argue that a dative constituent, which expresses a participant receiving a thing in a transfer event, is located somewhere on a continuum between a complement and a modifier. A recipient is sometimes analyzed as a semantically crucial participant for a transfer event to be semantically complete. This is because the transfer event is usually analyzed as consisting of three crucial participants, namely, a giver, a thing given and a recipient. However, the recipient is in some contexts perceived as not as semantically crucial as the other two participants as in John donates blood every month. On the other hand, the recipient in John gave an expensive birthday present to his mother, can be perceived to be a semantically crucial participant. That means the recipient can be perceived as a complement in some contexts and as a modifier in some others. Since the dative PP denoting a recipient fluctuates on the complement-modifier continuum, it is not surprising that the dative PP in Mandarin Chinese can occur both preverbally and postverbally according to the headcomplement and modifier-head constituent orders in Mandarin Chinese. However, Sanders and Uehara (2012) found that the dative [g\u011bi+NP] occur only preverbally in a speech corpus of Beijing Mandarin Chinese. This fact may suggest that the dative [g\u011bi+NP] in spoken Beijing Mandarin Chinese is perceived to be modifier-like rather than complement-like. The examples below illustrate the preverbal dative [g\u011bi+NP] . That is, the second verb is grammaticalized into a preposition marking a dative and benefactive NP. The PP functioning as a complement and a modifier occurs after the main VP. Therefore, the fact that the dative and benefactive [h\u00e2y+NP] constituents do not occur preverbally matches the predominant head-complement/modifier constituent order in Thai.",
	"cite_spans": [
	{
	"start": 2563,
	"end": 2571,
	"text": "[g\u011bi+NP]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "The fourth question is why the causative g\u011bi is not productive in Mandarin Chinese. Unlike the benefactive g\u011bi and the dative g\u011bi, which are grammaticalized into prepositions, the causative g\u011bi is more verb-like in that it can be negated. Notice that the causative g\u011bi appears in the same position as the benefactive g\u011bi, i.e. the preverbal position, which bears two consequences. The first consequence is that the preverbal g\u011bi tends to be analyzed as the benefactive marker functioning as a modifier of the main VP, which corresponds to the predominant modifier-head constituent order in Mandarin Chinese, rather than as the causative verb. The second consequence is that the preverbal g\u011bi in some cases can give rise to ambiguity between the causative and the benefactive readings. It is found that the other causative verbs r\u00e0ng and ji\u00e0o are used more frequently than g\u011bi in order to avoid ambiguity as stated earlier in the paper.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "The last question is why g\u011bi is not used as a clause subordinator in Mandarin Chinese but h\u00e2y is in Thai? Moreover, why is the clause subordinator h\u00e2y used highly productively in Thai? A complex construction consists of a matrix clause and a subordinating clause. Most subordinating clauses function as modifiers of the matrix VPs. In Mandarin Chinese, modifiers precede heads. Therefore, the postverbal position is not a perfect site for a verb to be grammaticalized into a subordinator in Mandarin Chinese. This is the reason why we do not find the postverbal subordinator g\u011bi in Mandarin Chinese. In contrast, the postverbal position is a perfect site for a verb to be grammaticalized into a subordinator introducing a subordinating clause in Thai because it matches the head-modifier constituent order in the language. That is why h\u00e2y is used as subordinator with a high degree of productivity in Thai.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "However, it is noted in some previous works that g\u011bi is used as a subordinator to introduce an adverbial clause occurring after a matrix clause in the head-adverbial clause order. This use of g\u011bi is exemplified by (35). Another observation can be made regarding the grammaticalized passive marker g\u011bi in Mandarin Chinese. It is noted in Thepkanjana and Uehara (2008) that the passive g\u011bi in the structural schema [g\u011bi + NP + VP] has been developed into what Newman (1993b: 477) calls \"the prefixal g\u011bi in passive constructions\" as in (35). From Newman (1993b: 477) (37) t\u0101 g\u011bi-m\u00e0-le he PASSIVE-scold-ASP 'He/She was scolded.'",
	"cite_spans": [
	{
	"start": 337,
	"end": 366,
	"text": "Thepkanjana and Uehara (2008)",
	"ref_id": "BIBREF49"
	},
	{
	"start": 458,
	"end": 477,
	"text": "Newman (1993b: 477)",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "This phenomenon, which indicates that the second verb becomes the head which the prefix g\u011bi is attached to, corresponds with the modifierhead pattern constituent order in Mandarin Chinese.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Effects of constituent or der s on pat ter ns of gr ammat icalizat ionof h\u00e2y an d g\u011bi",
	"sec_num": "3."
	},
	{
	"text": "This paper presents commonalities and differences in the grammaticalization of h\u00e2y in Thai and g\u011bi in Mandarin Chinese and argues how different constituent orders in Thai and Mandarin Chinese bear on patterns of Grammaticalization of the two verbs in the two languages. It is found that the common functions shared by h\u00e2y and g\u011bi are (1) the ditransitive main use, (2) the dative-marking use, (3) the benefactive-marking use and (4) the causativemarking use. As for differences, h\u00e2y, not g\u011bi, is used as a subordinator connecting two clauses in a complex construction whereas g\u011bi, not h\u00e2y, is used as a passive marker. Five questions are posed regarding different patterns of grammaticalization of h\u00e2y and g\u011bi in Thai and Mandarin Chinese. Facts about different patterns of grammaticalization of the two morphemes under discussion are accounted for in terms of different constituent orders in Thai and Mandarin Chinese, i.e. head-modifier/complement in Thai, modifier-head and head-complement in Mandarin Chinese. It is argued that the headmodifier constituent order in Thai seems to be compatible with postverbal grammaticalized morphemes whereas the modifier-head order in Mandarin Chinese seems to be compatible with preverbal grammaticalized ones.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "PACLIC-27",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "",
	"sec_num": null
	}
	],
	"back_matter": [
	{
	"text": "This research work is partially supported by the Ratchadaphiseksomphot Endowment Fund of Chulalongkorn University (RES560530179-HS) awarded to the first author and a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (No. 24520416) awarded to the second author.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Acknowledgments",
	"sec_num": null
	},
	{
	"text": "provide us with key clues for solving data sparseness problems, which are inherent in the unsupervised WSD. The proposed model shows 91.14% average accuracy, which is 26.95% higher than the best performance obtained by a supervised method (Lesk's dictionary-based WSD). Our model obtains also a higher accuracy for English and Chinese, using Princeton WordNet and HowNet. ",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "annex",
	"sec_num": null
	}
	],
	"bib_entries": {
	"BIBREF29": {
	"ref_id": "b29",
	"title": "Criterial Features in L2 English",
	"authors": [
	{
	"first": "J",
	"middle": [
	"A"
	],
	"last": "Hawkins",
	"suffix": ""
	},
	{
	"first": "L",
	"middle": [],
	"last": "Filipovi\u0107",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Hawkins, J.A. & Filipovi\u0107, L. (2012). Criterial Features in L2 English. Cambridge: Cambridge University Press.",
	"links": null
	},
	"BIBREF30": {
	"ref_id": "b30",
	"title": "Developing corpus-based word lists for English language learning and teaching: A critical appraisal of the English Vocabulary Profile",
	"authors": [
	{
	"first": "Y",
	"middle": [],
	"last": "Tono",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Tono, Y. 2012a. Developing corpus-based word lists for English language learning and teaching: A critical appraisal of the English Vocabulary Profile. In J. Thomas & A.",
	"links": null
	},
	"BIBREF31": {
	"ref_id": "b31",
	"title": "Input, Process and Product: Developments in Teaching and Language Corpora",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Boulton",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "314--328",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Boulton (eds). Input, Process and Product: Developments in Teaching and Language Corpora (pp.314-328). Brno: Masaryk University Press.",
	"links": null
	},
	"BIBREF32": {
	"ref_id": "b32",
	"title": "International Corpus of Crosslinguistic Interlanguage: Project overview and a case study on the acquisition of new verb co-occurrence patterns",
	"authors": [
	{
	"first": "Y",
	"middle": [],
	"last": "Tono",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "Developmental and Crosslinguistic Perspectives in Learner Corpus Research",
	"volume": "",
	"issue": "",
	"pages": "27--46",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Tono, Y. 2012b. International Corpus of Crosslinguistic Interlanguage: Project overview and a case study on the acquisition of new verb co-occurrence patterns. In Y. Tono, Y. Kawaguchi & M. Minegishi (eds.) Developmental and Crosslinguistic Perspectives in Learner Corpus Research (pp.27-46). Amsterdam: John Benjamins.",
	"links": null
	},
	"BIBREF33": {
	"ref_id": "b33",
	"title": "The CEFR-J: Adapting the CEFR for English language teaching in Japan",
	"authors": [
	{
	"first": "Y",
	"middle": [],
	"last": "Tono",
	"suffix": ""
	},
	{
	"first": "M",
	"middle": [],
	"last": "Negishi",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "JALT Framework & Language Portfolio SIG Newsletter",
	"volume": "",
	"issue": "8",
	"pages": "5--12",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Tono, Y. & Negishi, M. 2012. The CEFR-J: Adapting the CEFR for English language teaching in Japan. JALT Framework & Language Portfolio SIG Newsletter No.8 (September, 2012), pp. 5-12.",
	"links": null
	},
	"BIBREF34": {
	"ref_id": "b34",
	"title": "Linggle: a web-scale language reference search engine",
	"authors": [
	{
	"first": "References",
	"middle": [],
	"last": "Chang",
	"suffix": ""
	},
	{
	"first": "Jason",
	"middle": [
	"S"
	],
	"last": "",
	"suffix": ""
	}
	],
	"year": 2008,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "References Chang, Jason. S. 2008. Linggle: a web-scale language reference search engine. Unpublished manuscript.",
	"links": null
	},
	"BIBREF35": {
	"ref_id": "b35",
	"title": "Corpus analysis of the world wide web",
	"authors": [
	{
	"first": "William",
	"middle": [
	"H"
	],
	"last": "Fletcher",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "The Encyclopedia of Applied Linguistics",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Fletcher, William H. 2012. Corpus analysis of the world wide web.\" In The Encyclopedia of Applied Linguistics.",
	"links": null
	},
	"BIBREF36": {
	"ref_id": "b36",
	"title": "Word sketch: Extraction and display of significant collocations for lexicography",
	"authors": [
	{
	"first": "Adam",
	"middle": [],
	"last": "Kilgarriff",
	"suffix": ""
	},
	{
	"first": "David",
	"middle": [],
	"last": "Tugwell",
	"suffix": ""
	}
	],
	"year": 2001,
	"venue": "Proceedings of COLLOCTION: Computational Extraction, Analysis and Exploitation workshop",
	"volume": "",
	"issue": "",
	"pages": "32--38",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Kilgarriff, Adam, and David Tugwell. 2001. Word sketch: Extraction and display of significant collocations for lexicography. In Proceedings of COLLOCTION: Computational Extraction, Analysis and Exploitation workshop, pp. 32-38.",
	"links": null
	},
	"BIBREF37": {
	"ref_id": "b37",
	"title": "Googleology is bad science",
	"authors": [
	{
	"first": "Adam",
	"middle": [],
	"last": "Kilgarriff",
	"suffix": ""
	}
	],
	"year": 2007,
	"venue": "Computational linguistics",
	"volume": "33",
	"issue": "1",
	"pages": "147--151",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Kilgarriff, Adam. 2007. Googleology is bad science.\" Computational linguistics 33(1), pp. 147-151.",
	"links": null
	},
	"BIBREF38": {
	"ref_id": "b38",
	"title": "Concept discovery from text",
	"authors": [
	{
	"first": "Dekang",
	"middle": [],
	"last": "Lin",
	"suffix": ""
	},
	{
	"first": "Patrick",
	"middle": [],
	"last": "Pantel",
	"suffix": ""
	}
	],
	"year": 2002,
	"venue": "Proceedings of COLING",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Lin, Dekang, and Patrick Pantel. 2002. Concept discovery from text. In Proceedings of COLING.",
	"links": null
	},
	"BIBREF39": {
	"ref_id": "b39",
	"title": "New tools for web-scale n-grams",
	"authors": [
	{
	"first": "Dekang",
	"middle": [],
	"last": "Lin",
	"suffix": ""
	},
	{
	"first": "Kenneth",
	"middle": [],
	"last": "Church",
	"suffix": ""
	},
	{
	"first": "Heng",
	"middle": [],
	"last": "Ji",
	"suffix": ""
	},
	{
	"first": "Satoshi",
	"middle": [],
	"last": "Sekine",
	"suffix": ""
	},
	{
	"first": "David",
	"middle": [],
	"last": "Yarowsky",
	"suffix": ""
	},
	{
	"first": "Shane",
	"middle": [],
	"last": "Bergsma",
	"suffix": ""
	},
	{
	"first": "Kailash",
	"middle": [],
	"last": "Patil",
	"suffix": ""
	}
	],
	"year": 2010,
	"venue": "Proceedings of LREC",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Lin, Dekang, Kenneth Church, Heng Ji, Satoshi Sekine, David Yarowsky, Shane Bergsma, Kailash Patil et al. 2010. \"New tools for web-scale n-grams.\" Proceedings of LREC.",
	"links": null
	},
	"BIBREF40": {
	"ref_id": "b40",
	"title": "Using Web N-Grams to Help Second-Language Speakers",
	"authors": [
	{
	"first": "Martin",
	"middle": [],
	"last": "Potthast",
	"suffix": ""
	},
	{
	"first": "Martin",
	"middle": [],
	"last": "Trenkmann",
	"suffix": ""
	},
	{
	"first": "Benno",
	"middle": [],
	"last": "Stein",
	"suffix": ""
	}
	],
	"year": 2010,
	"venue": "Proceedings of SIGIR Web N-Gram Workshop",
	"volume": "",
	"issue": "",
	"pages": "49--49",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Potthast, Martin, Martin Trenkmann, and Benno Stein. Using Web N-Grams to Help Second-Language Speakers. 2010. In Proceedings of SIGIR Web N-Gram Workshop, pages 49-49.",
	"links": null
	},
	"BIBREF41": {
	"ref_id": "b41",
	"title": "Utilizing lexical data from a Web-derived corpus to expand productive collocation knowledge",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Wu",
	"suffix": ""
	},
	{
	"first": "Ian",
	"middle": [
	"H"
	],
	"last": "Shaoqun",
	"suffix": ""
	},
	{
	"first": "Margaret",
	"middle": [],
	"last": "Witten",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Franken",
	"suffix": ""
	}
	],
	"year": 2010,
	"venue": "ReCALL",
	"volume": "22",
	"issue": "1",
	"pages": "83--102",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Wu, Shaoqun, Ian H. Witten, and Margaret Franken. 2010. Utilizing lexical data from a Web-derived corpus to expand productive collocation knowledge.\" ReCALL 22(1), pp. 83-102.",
	"links": null
	},
	"BIBREF42": {
	"ref_id": "b42",
	"title": "Making preferences more active",
	"authors": [
	{
	"first": "Yorick",
	"middle": [],
	"last": "Wilks",
	"suffix": ""
	}
	],
	"year": 1978,
	"venue": "Artificial Intelligence",
	"volume": "11",
	"issue": "3",
	"pages": "197--223",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Wilks, Yorick. 1978. Making preferences more active. Artificial Intelligence 11(3), pp. 197-223.",
	"links": null
	},
	"BIBREF43": {
	"ref_id": "b43",
	"title": "The Status of Marker Gei in Mandarin Chinese",
	"authors": [
	{
	"first": "Dan",
	"middle": [],
	"last": "Xu",
	"suffix": ""
	}
	],
	"year": 1994,
	"venue": "Journal of Chinese Linguistics",
	"volume": "22",
	"issue": "2",
	"pages": "363--394",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Dan Xu. 1994. The Status of Marker Gei in Mandarin Chinese. Journal of Chinese Linguistics, 22(2): 363-394.",
	"links": null
	},
	"BIBREF44": {
	"ref_id": "b44",
	"title": "Give' Constructions in Malay, Thai and Mandarin Chinese: A Polygrammaticization Perspective",
	"authors": [
	{
	"first": "Ha",
	"middle": [],
	"last": "Foong",
	"suffix": ""
	},
	{
	"first": "Shoichi",
	"middle": [],
	"last": "Yap",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Iwasaki",
	"suffix": ""
	}
	],
	"year": 1998,
	"venue": "Proceedings of the 34 th Annual Meeting of the Chicago Linguistic Society",
	"volume": "",
	"issue": "",
	"pages": "421--438",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Foong Ha Yap and Shoichi Iwasaki. 1998. 'Give' Constructions in Malay, Thai and Mandarin Chinese: A Polygrammaticization Perspective. Proceedings of the 34 th Annual Meeting of the Chicago Linguistic Society, 421-438.",
	"links": null
	},
	"BIBREF45": {
	"ref_id": "b45",
	"title": "From Causative to Passive: A Passage in Some East and Southeast Asian Languages",
	"authors": [
	{
	"first": "Ha",
	"middle": [],
	"last": "Foong",
	"suffix": ""
	},
	{
	"first": "Shoichi",
	"middle": [],
	"last": "Yap",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Iwasaki",
	"suffix": ""
	}
	],
	"year": 2003,
	"venue": "Cognitive Linguistics and Non-Indo-European Languages. Mouton de Gruyter",
	"volume": "",
	"issue": "",
	"pages": "419--445",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Foong Ha Yap and Shoichi Iwasaki. 2003. From Causative to Passive: A Passage in Some East and Southeast Asian Languages. In Eugene H. Casad and Gary B. Palmer (eds.) Cognitive Linguistics and Non-Indo-European Languages. Mouton de Gruyter, Berlin & New York, 419-445.",
	"links": null
	},
	"BIBREF46": {
	"ref_id": "b46",
	"title": "A Cognitive Grammar Approach to Mandarin Gei",
	"authors": [
	{
	"first": "John",
	"middle": [],
	"last": "Newman",
	"suffix": ""
	}
	],
	"year": 1993,
	"venue": "Journal of Chinese Linguistics",
	"volume": "21",
	"issue": "2",
	"pages": "313--336",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "John Newman. 1993a. A Cognitive Grammar Approach to Mandarin Gei. Journal of Chinese Linguistics, 21(2): 313-336.",
	"links": null
	},
	"BIBREF47": {
	"ref_id": "b47",
	"title": "The Semantics of Giving in Mandarin",
	"authors": [
	{
	"first": "John",
	"middle": [],
	"last": "Newman",
	"suffix": ""
	}
	],
	"year": 1993,
	"venue": "Conceptualizations and Mental Processing in Language. Mouton de Gruyter",
	"volume": "",
	"issue": "",
	"pages": "433--485",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "John Newman. 1993b. The Semantics of Giving in Mandarin. In Richard A. Geiger and Brygida Rudzka-Ostyn (eds.) Conceptualizations and Mental Processing in Language. Mouton de Gruyter, Berlin & New York. 433-485.",
	"links": null
	},
	"BIBREF48": {
	"ref_id": "b48",
	"title": "Give: A Cognitive Linguistic Study",
	"authors": [
	{
	"first": "John",
	"middle": [],
	"last": "Newman",
	"suffix": ""
	}
	],
	"year": 1996,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "John Newman. 1996. Give: A Cognitive Linguistic Study. Mouton de Gruyter, Berlin & New York.",
	"links": null
	},
	"BIBREF49": {
	"ref_id": "b49",
	"title": "The Verb of Giving in Thai and Mandarin Chinese as a Case Study of Polysemy: A Comparative Study",
	"authors": [
	{
	"first": "Kingkarn",
	"middle": [],
	"last": "Thepkanjana",
	"suffix": ""
	},
	{
	"first": "Satoshi",
	"middle": [],
	"last": "Uehara",
	"suffix": ""
	}
	],
	"year": 2008,
	"venue": "Language Sciences",
	"volume": "30",
	"issue": "6",
	"pages": "621--651",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Kingkarn Thepkanjana and Satoshi Uehara. 2008. The Verb of Giving in Thai and Mandarin Chinese as a Case Study of Polysemy: A Comparative Study. Language Sciences, 30(6): 621-651.",
	"links": null
	},
	"BIBREF50": {
	"ref_id": "b50",
	"title": "The Grammaticalization of Passive Morphology",
	"authors": [
	{
	"first": "Martin",
	"middle": [],
	"last": "Haspelmath",
	"suffix": ""
	}
	],
	"year": 1990,
	"venue": "Studies in Language",
	"volume": "14",
	"issue": "1",
	"pages": "25--72",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Martin Haspelmath. 1990. The Grammaticalization of Passive Morphology. Studies in Language, 14(1): 25-72.",
	"links": null
	},
	"BIBREF51": {
	"ref_id": "b51",
	"title": "Syntax: Structure, Meaning and Function",
	"authors": [
	{
	"first": "Robert",
	"middle": [
	"D"
	],
	"last": "Van Valin",
	"suffix": ""
	},
	{
	"first": "Jr",
	"middle": [],
	"last": "",
	"suffix": ""
	},
	{
	"first": "Randy",
	"middle": [
	"J"
	],
	"last": "Lapolla",
	"suffix": ""
	}
	],
	"year": 1997,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Robert D. van Valin, Jr. and Randy J. LaPolla. 1997. Syntax: Structure, Meaning and Function. Cambridge University Press, Cambridge.",
	"links": null
	},
	"BIBREF52": {
	"ref_id": "b52",
	"title": "A Syntactic Classification of the Synchronic Use of G\u011bi in Beijing Mandarin: A Spoken Corpus-based Case Study of its Polyfunctionality",
	"authors": [
	{
	"first": "M",
	"middle": [],
	"last": "Robert",
	"suffix": ""
	},
	{
	"first": "Satoshi",
	"middle": [],
	"last": "Sanders",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Uehara",
	"suffix": ""
	}
	],
	"year": 2012,
	"venue": "Chinese Language and Discourse",
	"volume": "3",
	"issue": "2",
	"pages": "167--199",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Robert M. Sanders and Satoshi Uehara. 2012. A Syntactic Classification of the Synchronic Use of G\u011bi in Beijing Mandarin: A Spoken Corpus-based Case Study of its Polyfunctionality. Chinese Language and Discourse, 3(2): 167-199.",
	"links": null
	},
	"BIBREF53": {
	"ref_id": "b53",
	"title": "Foundations of Cognitive Grammar, volume I. Theoretical Prerequisites",
	"authors": [
	{
	"first": "Ronald",
	"middle": [
	"W"
	],
	"last": "Langacker",
	"suffix": ""
	}
	],
	"year": 1987,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Ronald W. Langacker. 1987. Foundations of Cognitive Grammar, volume I. Theoretical Prerequisites. Stanford University Press, Stanford, California.",
	"links": null
	},
	"BIBREF54": {
	"ref_id": "b54",
	"title": "An investigation of hafl y complex constructions in Thai",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Suda Rangkupan",
	"suffix": ""
	}
	],
	"year": 1997,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Suda Rangkupan. 1997. An investigation of hafl y complex constructions in Thai. Available from: <http://www.buffalo.edu/soc- sci/linguistics/people/students/ma_theses/rangkupa n/RANGKMA.PDF> PACLIC-27",
	"links": null
	}
	},
	"ref_entries": {
	"FIGREF0": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "Conference Keynote Speeches 1.Words and Rules Revisited: Reassessing the Role of Construction and Memory in Language ........................................................................................................ Alec Marantz 2. A Principle-Based Approach for Natural Language Processing: Eliminating the Shortcomings of Rule-Based Systems with Statistics ..................................... Wen-Lian Hsu 3. Extracting \"Criterial Features\" for the CEFR Levels Using Corpora of EFL Learners' Written Essays .................................................................................... Yukio Tono 4. It's about Time: More and More Sophisticated Statistical Methods in Corpus Linguistics .......................................................................................................... Stefan Th. Gries 5. Linking Text with Data and Knowledge Bases .................................................. Language Learners' Production Data for Understanding of L2 Learning Systems ............................................................................................. Yukio Tono 2. Introducing Linggle: From Concordance to Linguistic Search Engine ........... Jason S. Chang Main Conference Invited Talks 1. Statistical Machine Translation Based on Predicate-Argument Structure ......."
	},
	"FIGREF1": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "Kwon 3. Effects of Constituent Orders on Grammaticalization Patterns of the Serial Verbs for 'Give' in Thai and Mandarin Chinese .............................................. Kingkarn Thepkanjana and Satoshi Uehara PACLIC 27 Papers Oral Presentation Session 1A: Grammar and Syntax 1. Global Approach to Scalar Implicatures in Dynamic Semantics ..................... Jae-Il Yeom 2. Unification of Numeral Classifiers and Plural Markers: Empirical Facts and Implications......................................................................................................"
	},
	"FIGREF2": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "treat the systematic mapping between syntax and semantics of language (i.e. the compositional aspect of meaning).An increasing number of researchers have become interested in combining linguistic theories, which treat the compositional aspect of meaning, with statistical modeling of language.On the other hand, the community of knowledge-mining and semantic search has constructed large knowledge bases such as Freebase, Yago and Wikipedia. Although these knowledge-bases have been constructed independently of the interests in the NLP research community, they provide essential resources for research on Natural Language Understanding, which aims to develop a system which understands language as human being does. The first step of such an understanding system is to relate surface forms of language with corresponding units in knowledge-bases. Once text is mapped to representation in the knowledge domain, one can perform inferences of various sorts by combining it with knowledge in the knowledge base. Inferences, which combine information embedded within text with human knowledge which is external to text, are deemed essential in text understanding.The two streams of research in the above seem to be tackling the same problem of how surface expressions in text can be linked with extra-linguistic representation in the knowledge domain, and what roles the structure of language plays in such a linking process."
	},
	"FIGREF3": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "Paraphrasing and structures of sentences: While semantics of words have been studied extensively both in distributional semantics and traditional linguistics (e.g. synonyms, antonyms, etc.), semantics of larger units such as phrases and clauses have not been studied with similar degrees of details. Paraphrase recognition by structure alignment will provide a framework to capture semantics of larger units in language than words. We discuss how structures of sentences together with inferences based on meaning can give fine grained explanation of paraphrases, and how such research will contribute to the task of linking text with knowledge."
	},
	"FIGREF4": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "as selectional preference and widely reported in the literature. Linggle proposes preferred clusters of synonymous query arguments of adjectives, nouns and verbs. The clustering is achieved by building on Lin and Pantel (2002)'s large-scale repository of dependencies and word similarity scores and on an existing method for selectional preference induction with a Latent Dirichlet Allocation (LDA) model. , Chinese Academy of Sciences No. 95, Zhong Guan Cun East Road, Beijjing 100190, China cqzong@lpr.ia.ac.cn"
	},
	"FIGREF5": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "-Translation (ATT) framework, which is just based on the PAS and SC-PAS. As the following figure shows, this framework divides the whole translation process into three steps: (1) Analysis: to analyze the source sentences and obtain their PASs (or SC-PASs) automatically; (2) Transformation: to convert the source-side PASs (or SC-PASs) to target side by predicate-aware transformation rules;(3) Translation: this step is further divided into two parts: (a)element translation is to translate each element of PAS (or SC-PAS); (b)translation by global reordering is to combine the resulting translation candidates to translate the entire structure. By taking advantage of PAS (or SC-PAS), the ATT framework can well keep the semantic structure consistency of the source language and the target language and consequently show the great potential to improve translation quality."
	},
	"FIGREF6": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "(d) Even though the g\u011bi-marked dative PP in Mandarin Chinese is claimed by many researchers to occur both before and after the main VP, only the preverbal g\u011bi-marked dative PPs, not the postverbal ones, are attested in a Beijing Mandarin speech corpus (Sanders and Uehara, 2012). (e) The g\u011bi-marked benefactive PP in Mandarin Chinese can occur only before the main verb phrase. (f) The postverbal [h\u00e2y+NP] in Thai and the preverbal [g\u011bi+ NP] in Mandarin Chinese can be ambiguous between the dative and benefactive interpretations if the main VP incorporates the sense of giving. (g) The structural schemas of the causative and the passive g\u011bi are identical. (h) The causative use of h\u00e2y in Thai is productive but that of g\u011bi in Mandarin Chinese is not."
	},
	"FIGREF7": {
	"num": null,
	"uris": null,
	"type_str": "figure",
	"text": "The tree is very tall. (29) She read the newspaper in the living room. (30) She went to see a movie after dinner."
	},
	"TABREF0": {
	"type_str": "table",
	"content": "<table><tr><td>Masayuki Asahara, Sachi Yasuda, Hikari Konishi, Mizuho Imada, and Kikuo</td></tr><tr><td>Yongmei Tan and Eduard Hovy Maekawa</td></tr><tr><td>Oral Presentation Session 3A: Speech Perception 3. A Corpus-Hengbin Yan and Jonathan Webster</td></tr><tr><td>1. Towards a Yizhou Lan Oral Presentation Session 5A: Grammar and Syntax</td></tr><tr><td>Somnath Banerjee and Sivaji Bandyopadhyay 2. Oi Yee Kwong</td></tr><tr><td>Oral Presentation Session 2A: Discourse and Pragmatics 2. Hiromi Oyama, Mamoru Komachi, and Yuji Matsumoto</td></tr><tr><td>1. Li-chiung Yang</td></tr><tr><td>Wen-hui Sah 2. A Pin-Ning Tu and Shih-Ping Wang 2. Clausal-Dylan Marshall and Hamid Gomari</td></tr></table>",
	"text": "1. An Abstract Generation System for Social Scientific Papers .......................... Michio Kaneko and Dongli Han 2. Automatic Utterance Generation in Consideration of Nominatives and Emoticon Annotation .......................................................................................Yusuke Nishio, Mirai Miura, and Dongli Han3. Ensemble Approach for Fine-Grained Question Classification in Bengali ..... Prosodic Convergence, Divergence, and Feedback: Coherence and Meaning in Conversation .................................................................................................... Quantitative Comparative Study of Prosodic and Discourse Units, the Case of French and Taiwan Mandarin ......................................................................Laurent Pr\u00e9vot, Shu-Chuan Tseng, Alvin Cheng-Hsien Chen, and KlimPeshkov 3. Corpus-Based Research on Tense Analysis and Rhetorical Structure in Journal Article Abstracts............................................................................................. 1. A Novel Schema-Oriented Approach for Chinese New Word Identification Zhao Lu, Zhixian Yan, and Junzhong Gu 2. A Study of the Effectiveness of Suffixes for Chinese Word Segmentation ... Xiaoqing Li, Chengqing Zong, and Keh-Yih Su 3. Learning Fine-Grained Selectional Restrictions .......................................... Revised Motor Theory of L2 Speech Perception ......................... Difficulties in Perception and Pronunciation of Mandarin Chinese Disyllabic Word Tone Acquisition: A Study of Some Japanese University Students ..... Exploring the Chinese Mental Lexicon with Word Association Norms ........ Towards Automatic Error Type Classification of Japanese Language Learners' Writings .......................................................................................................... 1. The Development of Coherence in Narratives: Causal Relations ................. Packaging of Path of Motion in Second Language Acquisition of Russian and Spanish ...................................................................................... Kawai Chui, Hsiang-Lin Yeh, Wen-Chun Lan, and Yu-Han Cheng 3. Age Related Differences in Language Usage and Reading between English Monolinguals and Bilinguals ......................................................................... 1. Compiling a Corpus of Taiwanese Students' Spoken English ....................... Lan-fen Huang 2. BCCWJ-TimeBank: Temporal and Event Information Annotation on Japanese Text ................................................................................................................ Based Approach to Linguistic Function ......................................... 215",
	"html": null,
	"num": null
	},
	"TABREF2": {
	"type_str": "table",
	"content": "<table><tr><td>Oral Presentation Session 6A: Morphology and Syntax</td></tr><tr><td>1. Charles Lam</td></tr><tr><td>2. Tomokazu Takehisa</td></tr><tr><td>Oral Presentation Session 6B: Phonetics</td></tr><tr><td>1. Are Mandarin Sandhi Tone 3 and Tone 2 the Same or Different? The Results</td></tr><tr><td>of</td></tr><tr><td>Hai Zhao, Tianjiao Yin, and Jingyi Zhang</td></tr><tr><td>2. Yu-Chun Wang and Richard Tzong-Han Tsai</td></tr><tr><td>3. Effects of Parsing Errors on Pre-Reordering Performance for</td></tr><tr><td>Chinese-to-Dan Han, Pascual Mart\u00ednez-G\u00f3mez, Yusuke Miyao, Katsuhito Sudoh, and</td></tr><tr><td>Masaaki Nagata</td></tr></table>",
	"text": "1. Vietnamese to Chinese Machine Translation via Chinese Character as Pivot ........................................................................................................................ Transliteration Extraction from Classical Chinese Buddhist Literature Using Conditional Random Fields ........................................................................... Japanese SMT ............................................................................. Reduplication across Categories in Cantonese .............................................. Yet Another Piece of Evidence for the Common Base Approach to Japanese Causative/Inchoative Alternations ................................................................. Functional Data Analysis ...........................................................................",
	"html": null,
	"num": null
	},
	"TABREF3": {
	"type_str": "table",
	"content": "<table><tr><td>Workshop Session 3 3. Hayeon Jang, Munhyong Kim, and Hyopil Shin Poster Session 1 Qiuping Huang, Derek F. Wong, Lidia S. Chao, Xiaodong Zeng, and Liangye He 5. Tonal Patterns in the 15th Century: A Corpus-Chihkai Lin Alec Marantz 1. A Takahiro Ohno, Zyunitiro Edani, Ayato Inoue, and Dongli Han New York University, USA 2. PADS Restoration and Its Importance in Reading Comprehension and marantz@nyu.edu</td></tr><tr><td>7. Rahul Sharma, Soma Paul, Riyaz Ahmad Bhat, and Sambhav Jain 1. Cross-Qingliang Miao, Huayu Lu, Shu Zhang, and Yao Meng 6. Nalinee Intasaw and Wirote Aroonmanakun Shian-jung Chen</td></tr><tr><td>8. Myths in Korean Morphology and Their Computational Implications ......... 2. Akira Ohtani Hee-Rahk Chae</td></tr><tr><td>Workshop Session 1</td></tr><tr><td>1. Chin-chi Chao</td></tr><tr><td>2. iPad Reading: Hsin-Chou Huang</td></tr><tr><td>3. A Generic Cognitively Motivated Web-Environment to Help People to</td></tr><tr><td>Liyin Chen and Siaw-Fong Chung Shinya Kunii and Hiroyuki Shinnou Become Quickly Fluent in a Michael Zock, Guy Lapalme, and Lih-Juang Fang</td></tr><tr><td>2. Spatial Particles in English: A Quantitative Corpus-Based Approach to the</td></tr><tr><td>Conceptualization of Alvin Cheng-Hsien Chen Workshop Session 2</td></tr><tr><td>Tri Thanh Nguyen 1. Evaluation on Second Language Collocational Congruency with 3. Qian Yong 9. ML-Eckhard Bick Ching-Ying Lee and Chih-cheng Lin</td></tr><tr><td>Oral Presentation Session 7B: Sentiment Analysis Poster Session 2 2. A Corpus-Ping-Yu Huang, Chien-Ming Chen, Nai-Lung Tsao, and David Wible 1. Yasuhiro Noguchi, Satoru Kogure, Makoto Kondo, Hideki Asoh, Ichiro Kobayashi, Akira Takagi, Tatsuhiro Konishi, and Yukihiro Itoh 1. Takafumi Suzuki, Erina Kanou, and Yui Arakawa 3. Zhao-Ming Gao</td></tr><tr><td>Po-Ya Angela Wang</td></tr></table>",
	"text": "Show in Academic Discourse........................................................................................................ Symmetry in Bodily Orientation ................................. Typological Stage of Counterfactuals in Mandarin ....................................... Event Sequence Model for Semantic Analysis of Time and Location in Dialogue System ............................................................................................ Collective Sentiment Classification Based on User Leniency and Product Popularity ....................................................................................................... Wenliang Gao, Naoki Yoshinaga, Nobuhiro Kaji, and Masaru Kitsuregawa 4. KOSAC: A Full-Fledged Korean Sentiment Analysis Corpus ...................... Lingual Link Discovery between Chinese and English Wiki Knowledge Bases .............................................................................................................. Locative Postpositions and Conceptual Structure in Japanese ...................... ................................................................................... ............................. Tuned Constraint Grammars ................................................................. Comparative Analyses of Textual Contents and Styles of Five Major Japanese Newspapers .................................................................................................... Hui-Chuan Lu and Yu-Hsin Chu 4. Augmented Parsing of Unknown Word by Graph-Based Semi-Supervised Learning ......................................................................................................... Based Approach..................... 483 Basic Principles for Segmenting Thai EDUs ................................................. Automatic Clause Boundary Annotation in the Hindi Treebank ................... Creative Language Learning Projects with Digital Media ............................ An Innovative Approach to New Literacies ........................... New Language ................................................. Computational Semantic Similarity ............................................................... Based Tool for Exploring Domain-Specific Collocations in English ........................................................................................................................ Automatic Identification of English Collocation Errors Based on Dependency Relations ........................................................................................................ Japanese Learning Support System Matching Individual Abilities ............ Meaning Representation ................................................................................ 563",
	"html": null,
	"num": null
	},
	"TABREF4": {
	"type_str": "table",
	"content": "<table><tr><td>PACLIC-27</td></tr><tr><td>A of Information Science, Academia Sinica</td></tr><tr><td>hsu@iis.sinica.edu.tw</td></tr><tr><td>Abstract</td></tr><tr><td>In natural language processing, an important task is to recognize various linguistic</td></tr><tr><td>1</td></tr></table>",
	"text": "",
	"html": null,
	"num": null
	},
	"TABREF5": {
	"type_str": "table",
	"content": "<table><tr><td colspan=\"2\">Type of Corpora Name</td><td>Features</td></tr><tr><td>Input corpus</td><td>ELT materials corpus (to be</td><td>ELT course books</td></tr><tr><td/><td>completed)</td><td>Major textbooks that claim to be</td></tr><tr><td/><td/><td>CEFR-based</td></tr><tr><td colspan=\"3\">Interaction corpus Classroom observation data 30 hours secondary school ELT</td></tr><tr><td/><td/><td>classes</td></tr><tr><td>Output corpus</td><td colspan=\"2\">JEFLL Corpus (0.7 million) Written, secondary school, CEFR</td></tr><tr><td/><td/><td>level</td></tr><tr><td/><td>NICT JLE Corpus</td><td>Spoken, interview test scripts,</td></tr><tr><td/><td>(2 million)</td><td>1,280 participants, CEFR level</td></tr><tr><td/><td>ICCI</td><td>Written, primary & secondary</td></tr><tr><td/><td>(0.6 million)</td><td>school, 9000 samples, CEFR level</td></tr><tr><td/><td>GTECfS Corpus</td><td>Written, exam scripts, 30,000</td></tr><tr><td/><td>(to be comleted)</td><td>samples, CEFR level</td></tr><tr><td/><td>MEXT Corpus</td><td>S/W 2000 students randomly</td></tr></table>",
	"text": "shows the list of corpora to be used for the project: Corpora used for the project Three types of corpora have been either newly compiled or re-organised: input, interaction, and output corpora. For input corpora, major ELT publishers' CEFR-based course materials have been scanned and processed by OCR. For output corpora, major learner corpora for Japanese EFL learners, the JEFLL Corpus and the NICT JLE Corpus, have been selected, but for our project, the essays originally classified according to the school grades or oral proficiency test scores, have been re-classified according to the estimated CEFR levels assigned by",
	"html": null,
	"num": null
	},
	"TABREF6": {
	"type_str": "table",
	"content": "<table><tr><td>In this paper, It's</td><td>Linguistic features Total n. of words Total n. of sentences N. of VPs N. of clauses N. of T-units N. of dependent clauses 102.6 MeanDecreaseGini 440.3 134.8 277.2 182.4 121.3 N. of complex T-units 114.6 N. of complex nominals 210.2 Table2: Variable importance measured by Mean Decrease of Gini</td></tr></table>",
	"text": "I will report on the performance of different machine learning techniques, including random forest, support vector machine, decision tree (C4.5), and na\u00efve Bayes over CEFR-level classified texts and compare which programs produce the best result and useful additional information to evaluate the importance of criterial features. of California, Santa Barbara stgries@linguistics.ucsb.edu",
	"html": null,
	"num": null
	},
	"TABREF7": {
	"type_str": "table",
	"content": "<table><tr><td>PACLIC-27</td></tr><tr><td>413,300\" and \"rocky beach 16,800\". The n-gram search implementation is achieved</td></tr><tr><td>through filtering, re-indexing, and populating Web 1T ngram in a HBase database and</td></tr><tr><td>augmenting them with the most frequent PoS for words (without disambiguation)</td></tr></table>",
	"text": "University of Foreign Studies y.tono@tufs.ac.jpAbstractIn this workshop, I will share my experience in the field of learner corpus research (LCR). First I will define learner corpora in terms of its design criteria. Second, I will show how L2 learners' production data as corpora can be exploited to find linguistic Printed collocation dictionaries and reference tools based on compiled corpora offer limited coverage of word usage while knowledge of collocations is vital for the competent use of a language. We propose to address these limitations with a comprehensive system that truly aims at letting learners \"know a word by the company it keeps\". Linggle (linggle.com) is a broad coverage language reference tool for English as Second Language learners (ESL). The system is designed to access words in context under various forms. determiner followed by a important synonym and the keyword role (e. g., play a key part 15,900). A natural language interface is also available for users that would be less familiar to pattern based search. For example the question \"How can I describe a beach?\" would retrieve two word chunks with count such as \"sandy beach",
	"html": null,
	"num": null
	},
	"TABREF8": {
	"type_str": "table",
	"content": "<table><tr><td>PACLIC-27</td><td/><td/><td/><td/><td/><td/><td/><td/><td/></tr><tr><td>\u6b64</td><td>\u9879</td><td>\u8ba1\u5212</td><td>\u5c06</td><td>\u5bf9</td><td>\u52b3\u52a8</td><td>\u5927\u4f17</td><td>\u63d0\u4f9b</td><td>\u51cf\u7a0e</td><td/><td>\u4f18\u60e0</td></tr><tr><td/><td/><td/><td/><td/><td/><td colspan=\"2\">(1) Analysis</td><td/><td/></tr><tr><td>\u6b64</td><td>\u9879</td><td>\u8ba1\u5212</td><td>\u5c06</td><td>\u5bf9</td><td>\u52b3\u52a8</td><td>\u5927\u4f17</td><td>\u63d0\u4f9b</td><td>\u51cf\u7a0e</td><td/><td>\u4f18\u60e0</td></tr><tr><td>[</td><td>A0</td><td>]1</td><td colspan=\"2\">[AM-ADV]2 [</td><td>A2</td><td>]3</td><td>[ Pred ]4</td><td>[</td><td>A1</td><td>]5</td></tr><tr><td colspan=\"4\">Kingkarn Thepkanjana (3a) Element translation Chulalongkorn University</td><td colspan=\"3\">Source-side PAS(\u63d0\u4f9b)</td><td colspan=\"4\">(2) Transfomation Satoshi Uehara Tohoku University</td></tr><tr><td colspan=\"4\">[A0]1: this project / this plan / ... Phayathai Road, Bangkok 10330</td><td/><td/><td colspan=\"5\">Target-side-like PAS 41 Kawauchi, Aoba-ku, Sendai</td></tr><tr><td colspan=\"5\">[AM-ADV]2: will / ... THAILAND [A2]3: to public / to the working masses / ... [Pred]4: provide / to provide / ... [A1]5: tax concessions / ... Kingkarn.T@chula.ac.th</td><td/><td colspan=\"5\">X1 uehara@intcul.tohoku.ac.jp X2 X3 X5 980-8576 JAPAN X4</td></tr><tr><td/><td/><td/><td colspan=\"5\">(3b) Translation by global reordering</td><td/><td/></tr></table>",
	"text": "",
	"html": null,
	"num": null
	},
	"TABREF9": {
	"type_str": "table",
	"content": "<table><tr><td colspan=\"3\">Ditransitive verb use</td><td/></tr><tr><td>Thai:</td><td colspan=\"4\">[h\u00e2y + NP1 + NP2]</td></tr><tr><td/><td colspan=\"4\">(thing) (recipient)</td></tr><tr><td colspan=\"2\">(1) s\u01d2ms\u00e0k h\u00e2y</td><td>\u014b\u0259n</td><td/><td>s\u01d2mchay</td></tr><tr><td colspan=\"2\">Somsak give</td><td colspan=\"3\">money Somchay</td></tr><tr><td colspan=\"5\">'Somsak gave Somchay some money.'</td></tr><tr><td colspan=\"5\">Mandarin Chinese: [g\u011bi + NP1 + NP2]</td></tr><tr><td/><td/><td colspan=\"3\">(recipient) (thing)</td></tr><tr><td colspan=\"2\">(2) Zh\u0101ngs\u0101n</td><td>g\u011bi</td><td>L\u01d0s\u00ec</td><td>qi\u00e1n</td></tr><tr><td colspan=\"2\">Zhangsan</td><td>give</td><td>Lisi</td><td>money</td></tr><tr><td colspan=\"5\">'Zhangsan gave Lisi some money.'</td></tr><tr><td colspan=\"3\">Dative-marking use</td><td/></tr><tr><td>Thai:</td><td colspan=\"4\">[V + NP1 +h\u00e2y+ NP2]</td></tr><tr><td/><td colspan=\"2\">(thing)</td><td colspan=\"2\">(recipient)</td></tr><tr><td colspan=\"2\">(3) s\u01d2ms\u00e0k so$ \u014b</td><td>\u014b\u01ddn</td><td/><td>h\u00e2y s\u01d2mchay</td></tr><tr><td colspan=\"2\">Somsak send</td><td colspan=\"3\">money give Somchay</td></tr><tr><td colspan=\"5\">'Somsak sent some money to Somchay.'</td></tr><tr><td colspan=\"5\">Mandarin Chinese: 2 schemas Schema 1: postverbal g\u011bi [V + NP1 + g\u011bi + NP2]</td></tr><tr><td/><td colspan=\"2\">(thing)</td><td/><td>(recipient)</td></tr><tr><td colspan=\"2\">(4) Zh\u0101ngs\u0101n</td><td>j\u00ec-le</td><td/><td>y\u00ec f\u0113ng</td></tr><tr><td colspan=\"2\">Zhangsan x\u00ecn</td><td colspan=\"2\">send-ASP g\u011bi L\u01d0s\u00ec</td><td>one CLS</td></tr><tr><td>letter</td><td/><td>give</td><td>Lisi</td></tr><tr><td colspan=\"5\">'Zhangsan mailed a letter to Lisi.'</td></tr><tr><td colspan=\"3\">Schema 2: preverbal g\u011bi</td><td/></tr><tr><td/><td colspan=\"4\">[g\u011bi + NP1 + V + NP2]</td></tr><tr><td/><td colspan=\"2\">(recipient)</td><td/><td>(thing)</td></tr><tr><td colspan=\"2\">(5) Zh\u0101ngs\u0101n</td><td>g\u011bi</td><td>L\u01d0s\u00ec</td><td>m\u01cei</td></tr><tr><td colspan=\"2\">Zhangsan y\u00ec</td><td>give b\u011bn</td><td>Lisi sh\u016b</td><td>buy</td></tr><tr><td>one</td><td/><td>CLS</td><td colspan=\"2\">book</td></tr><tr><td colspan=\"5\">'Zhangsan bought a book for (and gave it to)</td></tr><tr><td>Lisi'</td><td/><td/><td/></tr></table>",
	"text": "The first common function between h\u00e2y and g\u011bi is the ditransitive main verb use. H\u00e2y and g\u011bi in this use co-occur with two NPs following each other in a row. The structural schemas of the ditransitive verbs h\u00e2y and g\u011bi and some examples of this use are given below. Notice that the semantic roles of NP1 and NP2 in Thai and Mandarin Chinese are different.",
	"html": null,
	"num": null
	},
	"TABREF18": {
	"type_str": "table",
	"content": "<table><tr><td>1. Why does the benefactive [g\u011bi+NP] occur</td></tr><tr><td>only in the preverbal position, not the</td></tr><tr><td>postverbal position, in Mandarin Chinese?</td></tr><tr><td>2. Unlike the benefactive [g\u011bi+NP], the dative</td></tr><tr><td>[g\u011bi+NP] occurs both preverbally and</td></tr><tr><td>postverbally in Mandarin Chinese. Why does</td></tr><tr><td>the dative [g\u011bi+NP] behave differently from</td></tr><tr><td>the benefactive [g\u011bi+NP]?</td></tr><tr><td>3. Why do the dative [h\u00e2y+NP] and the</td></tr><tr><td>benefactive [h\u00e2y+NP] not occur in the</td></tr><tr><td>preverbal position in Thai?</td></tr><tr><td>4. Why is the causative g\u011bi not productive in</td></tr><tr><td>Mandarin Chinese?</td></tr><tr><td>5. Why is g\u011bi not used as a clause subordinator</td></tr><tr><td>in Mandarin Chinese? In contrast, why is h\u00e2y</td></tr><tr><td>used as a clause subordinator in Thai?</td></tr><tr><td>Moreover, why is the clause subordinator h\u00e2y</td></tr><tr><td>used highly productively in Thai?</td></tr><tr><td>The first question is why the benefactive</td></tr><tr><td>[g\u011bi+NP] occurs only in the preverbal position,</td></tr><tr><td>not the postverbal position, in Mandarin Chinese.</td></tr><tr><td>In order to answer this question, we have to</td></tr><tr><td>understand the role of the benefactive PP in a</td></tr><tr><td>sentence. The benefactive PP in a sentence</td></tr><tr><td>serves as a modifier, rather than a complement,</td></tr><tr><td>of the main VP because it is peripheral and can</td></tr><tr><td>be omitted. It functions like an adverbial phrase</td></tr><tr><td>modifying the main VP. It merely adds an extra</td></tr><tr><td>piece of information regarding who benefits from</td></tr><tr><td>the agent's action. Therefore, the preverbal</td></tr><tr><td>benefactive [g\u011bi+NP] matches the modifier-head</td></tr><tr><td>constituent order in Mandarin Chinese. The</td></tr><tr><td>postverbal benefactive [g\u011bi+NP] would violate</td></tr><tr><td>this constituent order in the language.</td></tr></table>",
	"text": "The second question is why the dative[g\u011bi+NP] behaves differently from the benefactive[g\u011bi+NP] in Mandarin Chinese. That is, the dative [g\u011bi+NP] occurs both preverbally and postverbally whereas the benefactive[g\u011bi+NP]",
	"html": null,
	"num": null
	},
	"TABREF19": {
	"type_str": "table",
	"content": "<table><tr><td colspan=\"4\">Data from Sanders' and Uehara's personal</td></tr><tr><td colspan=\"2\">communication</td><td/><td/></tr><tr><td>(33) m\u00e9i</td><td>g\u011bi</td><td>n\u01d0</td><td>xi\u011b</td></tr><tr><td>not</td><td>give</td><td>you</td><td>write</td></tr><tr><td colspan=\"3\">'I haven't written to you.'</td><td/></tr><tr><td>(34) w\u01d2</td><td colspan=\"2\">g\u011bi n\u01d0men shu\u014d</td><td>ya</td></tr><tr><td>I</td><td colspan=\"2\">give you (pl.) say</td><td>PART.</td></tr><tr><td colspan=\"2\">'Let me tell you.'</td><td/><td/></tr><tr><td colspan=\"4\">The third question is why the dative and</td></tr><tr><td colspan=\"4\">benefactive [h\u00e2y+NP] do not occur preverbally</td></tr><tr><td colspan=\"4\">in Thai. In the grammaticalization process, a</td></tr><tr><td colspan=\"4\">string of [V1+NP1] + [V2+NP2] is reanalyzed</td></tr><tr><td colspan=\"3\">into [V+NP1] + [P+NP2]</td><td/></tr></table>",
	"text": "in spoken Beijing Mandarin Chinese.",
	"html": null,
	"num": null
	},
	"TABREF20": {
	"type_str": "table",
	"content": "<table><tr><td>(35) Zha# ngsa# n</td><td colspan=\"2\">chafl ng ge#</td><td>ge& & i</td></tr><tr><td>Zhangsan</td><td>sing</td><td>song</td><td>give</td></tr><tr><td>ta#</td><td>ti# ng</td><td/><td/></tr><tr><td>he/she</td><td>hear</td><td/><td/></tr><tr><td colspan=\"4\">'Zhangsan sang a song for him/her to hear.'</td></tr><tr><td colspan=\"4\">However, this construction is not attested in a</td></tr><tr><td colspan=\"4\">Beijing Mandarin speech corpus according to</td></tr><tr><td colspan=\"4\">Sanders and Uehara (2012). To express this</td></tr><tr><td colspan=\"4\">meaning, the benefactive ge& & i is used instead as in</td></tr><tr><td>(36).</td><td/><td/><td/></tr><tr><td>(36) Zha# ngsa# n</td><td>ge& & i</td><td>ta#</td><td>chafl ng</td></tr><tr><td>Zhangsan</td><td>give</td><td colspan=\"2\">he/she sing</td></tr><tr><td>ge#</td><td/><td/><td/></tr><tr><td>song</td><td/><td/><td/></tr><tr><td colspan=\"4\">'Zhangsan sang a song for him/her.'</td></tr></table>",
	"text": "The fact that the subordinator ge& & i is not found in spoken Beijing Mandarin Chinese confirms our hypothesis that the postverbal position is not a perfect site for ge& & i to be grammaticalized into a subordinator.",
	"html": null,
	"num": null
	}
	}
	}
	}