Thesis (Ph. D.)--University of California, San Diego, 2007. / Title from first page of PDF file (viewed June 21, 2007). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references (p. 197-213).

Thesis (Ph.D.)--Hong Kong University of Science and Technology, 2007. / Includes bibliographical references (leaves 223-230). Also available in electronic version.

Thesis (Ph. D.)--University of Washington, 2006. / Vita. Includes bibliographical references (leaves 305-315).

Mandarin Chinese has a very interesting and complex aspectual system. This thesis studies this system from a new perspective, a modified version of Smith's (1991) two-component theory. It is shown that the modifications developed in the thesis increase the explanatory power of the theory so that a clearer picture of the Mandarin Chinese system is obtained. Situation aspect is compositional in nature. It is determined by the interaction of the intrinsic nature of the verb, the kind of argument the verb takes and certain other phrases that may occur in the sentence. In this thesis, a three level model is developed to account for the compositional nature of aspect. It is argued that aspect composition processes take place at three different levels: (i) the lexical level through some lexical processes; (ii) the subcategorizational level through the interaction of the verb and its arguments and (iii) the post-subcategorizational level through the interaction of the core sentence and certain adverbial constituents. This model has a few advantages. First, it defines clearly what plays what role in situation aspect composition processes. Secondly, this model makes it possible to talk in clear terms of verb types and situation types. Actually this model replaces Smith's situation shifting with situation formation. The three-level model is supplemented by two new criteria for verb classifications and a further distinction between boundedness and telicity. The two new criteria are: (i) the (±result) feature that distinguishes Accomplishment verbs from Achievement verbs and (ii) the (±bounded) feature that distinguishes Activity verbs from Semelfactive verbs. The distinction between boundedness and telicity differentiates temporal boundaries and spatial boundaries, the former is referred to by the feature (±bounded), and the latter is referred to by the feature (±telic). In light of the modified version of Smith's two component theory, the Mandarin Chinese aspectual system is shown to fit in the big picture of the universal grammar. The perfective and the imperfective are the two basic viewpoints in Mandarin Chinese. The perfective is marked by le and guo, and the imperfective is marked by zai and zhe. However, these viewpoints have some language specific properties. The most outstanding special properties are displayed by le. Unlike the perfective in languages like English and French, the perfective marker le does not provide a final endpoint. It only emphasizes the occurrence of a situation as a whole. As a consequence, it requires that the situation it marks is either (+bounded) or (+telic). Guo provides an endpoint to situations and so it is compatible with any situation type either open-ended or closed. However, as it emphasizes the experiential meaning of a particular event, it is limited in use pragmatically. Zai emphasizes the progress of a situation, it is not sensitive to endpoint at all. Its function is close to the English progressive form and like the English progressive form, it fits in the general imperfective schema. Zhe imposes a static view to situations, and because of this property, zhe is selective about situation types. In Mandarin Chinese, there are some constructions that show certain syntactic and semantic constraints. These constraints are studied in light of the aspectual theory adopted and developed in the thesis. And it is argued that a top-down approach advocated by the Construction Grammar and the Head-driven Phrase Structure Grammar may well explain the form-meaning correspondences of these constructions. / Graduate

We introduce a number of techniques for establishing the deadlock freedom of concurrent systems. Our methods are based on the local analysis (or at worst a directed global analysis) of networks. We identify the relationships between these techniques and the range of their application within a framework of deadlock freedom types that we have defined. We also show that the problem of proving total correctness may be translated to one of proving deadlock freedom, with the consequence that our techniques for proving deadlock freedom may be utilised to effect a total correctness proof.

Patterns are used as building blocks for design and construction in many fields such as architecture, music, literature, etc. Researchers and practitioners in the information systems area have been exploring patterns and using them in system analysis and design. Patterns found in the analysis stage, when analysts create conceptual models to abstractly represent domain reality, are call business patterns or analysis patterns. Although various business patterns were proposed in previous studies, we found that business semantics were missing in these patterns. These business patterns failed to show functionalities that is essential to patterns in general. Most of these patterns were also not capable of describing business activities, the dynamic aspect of business. This study is conducted to address these issues. In this thesis, we provide a brief literature review on business patterns, and discuss the major problems we found in these studies. Then we introduce our research approach and the major outcomes. We propose a new definition of business patterns with business semantics, which enables us to recover the missing functionality in business patterns. We suggest the key elements to represent business patterns, and propose a two-level template (functional and operational) to describe these elements. Based on theR²M approach, we propose a modeling method with graphical notations to describe the operational level of patterns, where business activities can be modeled. Examples and a case study are provided in this thesis to demonstrate how to use the modeling method and how to use business patterns in practice. / Business, Sauder School of / Management Information Systems, Division of / Graduate

Huge amount of data flows around the Internet every second, but for the data to be useful at its destination, it must be presented in a way such that the target has little problem interpreting it. Current data exchange technologies may rearrange the structure of data to suit expectations at the target. However, there may be semantics behind data (e.g. knowing the title of a book can determine its #pages) that may be violated after data translation. These semantics are expressed as integrity constraints (IC) in a database. Currently, there is no guarantee that the exchanged data conforms to the target’s ICs. As a result, existing applications (e.g. user queries) that assume such semantics will no longer function correctly. Current constraint repair techniques deal with data after it has been translated; thus take no consideration of the integrity constraints at the source. Moreover, such constraint repair methods usually involve addition/deletion/modification of data, which may yield incomplete or false data. We consider the constraints of both source and target schemas; together with the mapping, we can efficiently detect which constraint is violated and suggest ways to correct the mappings. / Science, Faculty of / Computer Science, Department of / Graduate

En traitement automatique de la langue, les différentes étapes d'analyse usuelles ont tour à tour amélioré la façon dont le langage peut être modélisé par les machines. Une étape d'analyse encore mal maîtrisée correspond à l'analyse sémantique. Ce type d'analyse permettrait de nombreuses avancées, telles que de meilleures interactions homme-machine ou des traductions plus fiables. Il existe plusieurs structures de représentation du sens telles que PropBank, les AMR et FrameNet. FrameNet correspond à la représentation en cadres sémantiques dont la théorie a été décrite par Charles Fillmore. Dans cette théorie, chaque situation prototypique et les différents éléments y intervenant sont représentés de telle sorte que deux situations similaires soient représentées par le même objet, appelé cadre sémantique. Le projet FrameNet est une application de cette théorie, dans laquelle plusieurs centaines de situations prototypiques sont définies. Le travail que nous décrirons ici s'inscrit dans la continuité des travaux déjà élaborés pour prédire automatiquement des cadres sémantiques. Nous présenterons quatre systèmes de prédiction, chacun ayant permis de valider une hypothèse sur les propriétés nécessaires à une prédiction efficace. Nous verrons également que notre analyse peut être améliorée en fournissant aux modèles de prédiction des informations raffinées au préalable, avec d'un côté une analyse syntaxique dont les liens profonds sont explicités et de l'autre des représentations vectorielles du vocabulaire apprises au préalable. / In natural language processing, each analysis step has improved the way in which language can be modeled by machines. Another step of analysis still poorly mastered resides in semantic parsing. This type of analysis can provide information which would allow for many advances, such as better human-machine interactions or more reliable translations. There exist several types of meaning representation structures, such as PropBank, AMR and FrameNet. FrameNet corresponds to the frame semantic framework whose theory has been described by Charles Fillmore (1971). In this theory, each prototypical situation and each different elements involved are represented in such a way that two similar situations are represented by the same object, called a semantic frame. The work that we will describe here follows the work already developed for machine prediction of frame semantic representations. We will present four prediction systems, and each one of them allowed to validate another hypothesis on the necessary properties for effective prediction. We will show that semantic parsing can also be improved by providing prediction models with refined information as input of the system, with firstly a syntactic analysis where deep links are made explicit and secondly vectorial representations of the vocabulary learned beforehand.

This thesis investigates lexical and grammatical aspect in two unrelated languages, Sencoten(the Saanich dialect of Straits Salish) and Japanese. In particular, the main focus is on how various perfect readings are derived in the two languages, which show striking similarities in this respect. In Sencoten, a particle kwlh yields various readings depending on the situation aspect and viewpoint aspect of the predicate with which it occurs (Kiyota 2006b). These various readings include an inceptive reading, an on-going situation reading, and a completion reading. The Japanese aspectual marker -tei- also induces a range of different readings: a progressive reading, a resultant state reading, and a perfect reading (Ogihara 1998a, Nishiyama 2006, a.o.). To account for these various readings, I propose that both kwlh in Sencoten and -tei- in Japanese are perfect markers. However, the actual semantic function of each is different: kwlh in Sencoten introduces a perfect time span (Pancheva 2003), whereas -tei- in Japanese denotes an anteriority relation between an event time and a reference time (Reichenbach 1947, Klein 1992, 1994), where the event time can be the time interval of a sub-event of a larger event. -Tei- also has a pragmatic component (or presupposition), just as Portner (2003) claims for the English perfect. Aspectual properties of predicates also play a crucial role in yielding the range of different readings. Therefore, this thesis also proposes a new aspectual classification of predicates in Sencoten and Japanese, which departs from the common classification of predicates based on Indo-European languages. In Sencoten, various readings are derived by interaction between the semantics of verbal predicates (i.e. lexical aspect), the semantics of the grammatical aspect (perfective or imperfective), and the semantics (and possibly pragmatics) of the perfect. In Japanese, the range of interpretations is due to interaction between the semantics of verbal aspect, the function of an adverb, and the semantics and pragmatics of the perfect. In other words, the same factors enter into my analysis of both Sencoten and Japanese, though there is one striking difference between the two languages: the perfective/imperfective opposition is involved in Sencoten, but not in Japanese. / Arts, Faculty of / Linguistics, Department of / Graduate

A database schema should consist of structures and semantic integrity constraints. Se mantic integrity constraints (SICs) are invariant restrictions on the static states of the stored data and the state transitions caused by the primitive operations: insertion, dele tion, or update. Traditionally, database design has been carried out on an ad hoc basis and focuses on structure and efficiency. Although the E-R model is the popular concep tual modelling tool, it contains few inherent SICs. Also, although the relational database model is the popular logical data model, a relational database in fourth or fifth normal form may still represent little of the data semantics. Most integrity checking is distributed to the application programs or transactions. This approach to enforcing integrity via the application software causes a number of problems. Recently, a number of systems have been developed for assisting the database design process. However, only a few of those systems try to help a database designer incorporate SICs in a database schema. Furthermore, current SIC representation languages in the literature cannot be used to represent precisely the necessary features for specifying declarative and operational semantics of a SIC, and no modelling tool is available to incorporate SICs. This research solves the above problems by presenting two models and one subsystem. The E-R-SIC model is a comprehensive modelling tool for helping a database designer in corporate SICs in a database schema. It is application domain-independent and suitable for implementation as part of an automated database design system. The SIC Repre sentation model is used to represent precisely these SICs. The SIC elicitation subsystem would verify these general SICs to a certain extent, decompose them into sub-SICs if necessary, and transform them into corresponding ones in the relational model. A database designer using these two modelling tools can describe more data semantics than with the widely used relational model. The proposed SIC elicitation subsystem can provide more modelling assistance for him (her) than current automated database design systems. / Business, Sauder School of / Graduate