This dissertation studies leftward dislocation of object Noun Phrases in Mandarin Chinese within the framework of Government and Binding theory. Although the canonical word order in Chinese is S(ubject)-V(erb)-O(bject), it also exhibits OSV and SOV word orders. After an introduction in Ch. 1, I discuss OSV constructions in Ch.2. I argue that the S-initial object NP is moved there, since its association with a gap in the canonical object position obeys the subjacency condition. Based on several diagnostic tests, I propose that Chinese has two kinds of short-distance NP fronting: one is A’-movement and the other is A-movement. Adopting the Split Infl Hypothesis, I postulate a fully articulated clause structure for Chinese. In particular, I propose that the fronted NP in A-movement lands in [Spec AgrOP] as a kind of overt raising, while the one in A’-movement further leaves that spec position and is CP-adjoined. I also examine long-distance NP fronting, showing that it is invariably A’-movement. In Ch. 3, I investigate object shift, which yields SOV constructions. I argue that this syntactic process represents a type of A-movement, not A’-movement as concluded in previous studies. Specifically, I propose that the subject NP and the object NP in this construction overtly raise to [Spec AgrSP] and [Spec AgrOP] respectively. In Ch.4, I examine the interactions between an object wh-NP and dou, the adverb of universal quantification. I propose that wh-phrases, like indefinites, can be either presuppositional or existential. If they are within VP (i.e. remain postverbal), they are subject to existential closure and get an existential/interrogative reading. If, however, they are outside VP (i.e. shifted to the left of dou), they define the range of the quantifier dou and obtain the presuppositional/universal reading. The conclusion is that there is a strict correlation between the S-structure positions of the wh-phrase and its interpretations. The exhaustive list reading of the in-situ wh-object associated with the interrogative reading is derived from the fact that it is in the scope of dou at S-structure. In Ch. 5, I summarize the major findings of this dissertation and raise several issues for future research.

In this thesis we propose three new word association measures for multi-word term extraction. We combine these association measures with LocalMaxs algorithm in our extraction model and compare the results of different multi-word term extraction methods. Our approach is language and domain independent and requires no training data. It can be applied to such tasks as text summarization, information retrieval, and document classification. We further explore the potential of using multi-word terms as an effective representation for general web-page summarization. We extract multi-word terms from human written summaries in a large collection of web-pages, and generate the summaries by aligning document words with these multi-word terms. Our system applies machine translation technology to learn the aligning process from a training set and focuses on selecting high quality multi-word terms from human written summaries to generate suitable results for web-page summarization.

No description available.

No description available.

No description available.

It is recognized that the World Heritage Site (WHS) status is a strong brand with exceptional quality and excellent reputation that attracts tourists to visit. This study applies and adapts the brand knowledge model to examine local stakeholders’ understanding of the WHS status as a brand. A case study approach was applied and a WHS in China was selected as the case. In total, 13 interviewees from local government, private business owners and residents participated in the study. The study showed that the three local stakeholders were familiar with the WHS status and shared the importance of the WHS status as intended by the WHS program to tourists. However, local stakeholders over emphasized the economic importance of the WHS status, and conservation became a tool to fulfill economic benefits. The results challenged the standpoint of the WHS program and showed that the WHS status was not a strong brand.

Word-Sense Disambiguation (WSD), holds promise for many NLP applications requiring broad-coverage language understanding, such as summarization (Barzilay and Elhadad, 1997) and question answering (Ramakrishnan et al., 2003). Recent studies have also shown that WSD can benefit machine translation (Vickrey et al., 2005) and information retrieval (Stokoe, 2005). Much work has focused on the computational treatment of sense ambiguity, primarily using data-driven methods. The most accurate WSD systems to date are supervised and rely on the availability of sense-labeled training data. This restriction poses a significant barrier to widespread use of WSD in practice, since such data is extremely expensive to acquire for new languages and domains. Unsupervised WSD holds the key to enable such application, as it does not require sense-labeled data. However, unsupervised methods fall far behind supervised ones in terms of accuracy and ease of use. In this thesis we explore the reasons for this, and present solutions to remedy this situation. We hypothesize that one of the main problems with unsupervised WSD is its lack of a standard formulation and general purpose tools common to supervised methods. As a first step, we examine existing approaches to unsupervised WSD, with the aim of detecting independent principles that can be utilized in a general framework. We investigate ways of leveraging the diversity of existing methods, using ensembles, a common tool in the supervised learning framework. This approach allows us to achieve accuracy beyond that of the individual methods, without need for extensive modification of the underlying systems. Our examination of existing unsupervised approaches highlights the importance of using the predominant sense in case of uncertainty, and the effectiveness of statistical similarity methods as a tool for WSD. However, it also serves to emphasize the need for a way to merge and combine learning elements, and the potential of a supervised-style approach to the problem. Relying on existing methods does not take full advantage of the insights gained from the supervised framework. We therefore present an unsupervised WSD system which circumvents the question of actual disambiguation method, which is the main source of discrepancy in unsupervised WSD, and deals directly with the data. Our method uses statistical and semantic similarity measures to produce labeled training data in a completely unsupervised fashion. This allows the training and use of any standard supervised classifier for the actual disambiguation. Classifiers trained with our method significantly outperform those using other methods of data generation, and represent a big step in bridging the accuracy gap between supervised and unsupervised methods. Finally, we address a major drawback of classical unsupervised systems – their reliance on a fixed sense inventory and lexical resources. This dependence represents a substantial setback for unsupervised methods in cases where such resources are unavailable. Unfortunately, these are exactly the areas in which unsupervised methods are most needed. Unsupervised sense-discrimination, which does not share those restrictions, presents a promising solution to the problem. We therefore develop an unsupervised sense discrimination system. We base our system on a well-studied probabilistic generative model, Latent Dirichlet Allocation (Blei et al., 2003), which has many of the advantages of supervised frameworks. The model’s probabilistic nature lends itself to easy combination and extension, and its generative aspect is well suited to linguistic tasks. Our model achieves state-of-the-art performance on the unsupervised sense induction task, while remaining independent of any fixed sense inventory, and thus represents a fully unsupervised, general purpose, WSD tool.

Children with Specific Language Impairment (SLI) often present with difficulties in learning new words compared to age-matched children with typical language development. These difficulties may affect the acquisition, storage, or retrieval of new words. Word-learning deficits impact on children’s vocabulary development and impede their language and literacy development. Findings from a wide range of studies investigating word-learning in children with SLI demonstrated that semantic and phonological knowledge are crucial to the word-learning process. However, intervention studies designed to improve the word-learning abilities in children with SLI are sparse. The experiments described in this thesis addressed this need to understand the effects of interventions on word-learning abilities. Further, the thesis describes the first investigation of word-learning abilities of New Zealand school-aged children with SLI. Specifically, the following three broad questions are asked: 1. What are the word-learning skills of New Zealand school-aged children with SLI compared to children with typical language development and which underlying language skills influence word-learning? 2. What are the immediate and longer term effects of phonological awareness and semantic intervention on word-learning and language skills in children with SLI? 3. What are the error patterns of children with SLI compared to children with typical language development when learning to produce new words and do these patterns change following phonological awareness and semantic intervention? The first experiment compared the word-learning abilities of 19 school-aged children with SLI (aged 6;2 to 8;3) to age-matched children with typical language development and revealed that children with SLI presented with significant difficulties to produce and to comprehend new words. After repeated exposure, children with SLI caught up to the performances of children with typical language development in learning to comprehend new words, but not on production of new words. Correlation analyses demonstrated that there were no correlations between the word-learning skills and other language measures for children with SLI, whereas the word-learning abilities of children with typical language development were correlated to their phonological awareness, semantic, and general language skills. In the second experiment, it was investigated whether there were also qualitative differences during word-learning between children with and without SLI additionally to the quantitative differences as revealed in the first experiment. Children’s erroneous responses during the word-learning tasks were categorised into phonological, semantic, substitution or random errors. A comparison of the children’s error patterns revealed that children with SLI presented with a different error pattern and made significantly more random errors than children with typical language development. However, after repeated exposure, children with SLI demonstrated a similar error pattern as children without SLI. Furthermore, it was examined whether a specific combination of phonological and semantic cues facilitated children’s learning of new words or whether there were word-specific features that facilitated children’s word-learning. No facilitative word-specific features could be identified. Analysis revealed that there were no significant effects of cueing on learning new words, but specific patterns could be derived for children with SLI. Children with SLI learned to comprehend more words that were presented with two semantic cues or one phonological and one semantic cue and learned to produce more words that were presented with two phonological cues. In the third experiment, the effectiveness of a combined phonological awareness and semantic intervention to advance children’s word-learning abilities was examined. Nineteen children with SLI (same participants as in experiment 1) participated in this intervention study that implemented an alternating treatment group design with random assignment of the participants. Children in group A received phonological awareness intervention followed by semantic intervention, whereas children in group B received the same interventions in the reverse order. Children’s word-learning abilities were assessed at pre-test, prior to the intervention, at mid-test after intervention phase 1, and at post-test, immediately following the completion of the second intervention phase. Each intervention itself was effective in significantly improving children’s fast mapping skills, however, gains in children’s word-learning abilities were only found for children in group A for production of new words. Extending the findings of the intervention effectiveness of phonological awareness and semantic intervention on word-learning as reported in experiment 3, it was investigated in experiment 4, whether the implemented intervention additionally influenced the error patterns of children with SLI. The erroneous responses of children with SLI on all word-learning probes at pre-, mid-, and post-test were categorised into the same error groups as described in the second experiment (semantic, phonological, substitution, and random errors). The error analyses revealed that children’s error profiles changed during the course of intervention and treatment specific effects on children’s erroneous responses were found. Post-intervention, children who received phonological awareness followed by semantic intervention displayed the same error patterns as children with typical language development, whereas children who received the same interventions in the reverse order maintained the same error pattern as displayed at pre-test. The final experiment examined the longer-term effects of the combined phonological awareness and semantic intervention reported in experiment 3 on the language and literacy development of children with SLI. Eighteen of the 19 children with SLI, who received the intervention reported in experiment 3, were available for re-assessment 6 months after the completion of the intervention. The children (aged 7;1 to 9;2 years) were re-assessed on a range of standardised and experimental measures. Data analysis revealed that 6 months post-intervention, all children were able to maintain their gains in phonological awareness, semantic, and decoding skills as displayed immediately after the intervention. Children’s general language and reading skills significantly improved following the intervention; however, children who received phonological awareness intervention followed by semantic intervention displayed significantly better reading outcomes than the children who received the same interventions in the reverse order. This thesis revealed that a combination of phonological awareness and semantic intervention can enhance the word-learning abilities of children with SLI. The combined intervention approach was also effective in additionally improving children’s general language skills and the reading of single non-words and real words, as well as connected text. The immediate and longer-term intervention effects provide evidence that advancing the semantic and phonological awareness skills is an effective intervention approach to support children with SLI in their word-learning and to furthermore promote their language and literacy development. However, the order of the implemented interventions played a significant role: Children in the current study profited most when they received phonological awareness intervention first, followed by semantic intervention.

Research has shown that cultural differences and the lack of experiences in the lives of young children can affect the rate of vocabulary development. In particular, children from different ability, socioeconomic status, and culturally and linguistically diverse groups are considered at risk for later academic achievement because their home experiences and word usage may be incongruent with that of the mainstream school cultural environment. Therefore, it has been suggested that to decrease the gap between children in need of vocabulary development and their typically achieving peers, instruction in vocabulary should systematically provide information about words and their uses. Thus, the purpose of this study was to examine the effect of a systematic vocabulary instructional technique in children with clinically depressed vocabulary skills. An additional goal was to examine the role of book type in the retention of novel vocabulary words among young African American children. Using an Adapted Alternating Treatments Design, five children were read two storybooks in the context of robust vocabulary training. Storybooks were used as a source for contextualizing novel vocabulary words. One book depicted an African American theme and images and the other depicted a Caucasian theme and images. Robust vocabulary instruction consisted of frequent and varied opportunities for word usage in meaningful contexts that stressed the relations between target words and previously acquired vocabulary. Childrens productive definitions were used to assess developing word knowledge at 4 periodic probes. Definitions were scored using a 4-stage continuum ranging from no knowledge to full concept knowledge. Results showed significant gains in word learning for novel words two weeks following conclusion of the study. The difference in scores between the instructional and control word sets resulted in a large effect size attributable to robust vocabulary instruction. African American children appeared to learn words at a deeper level from a storybook that displayed sociocultural images and experiences different from their own.

This project sought to find a rigorous and manageable method for measuring the difficulty of texts in te reo Māori written for children, beyond junior reading material in Māori-medium educational settings. The project examined a range of readability measures based on semantic and/or syntactic features of text, following the work of Warwick Elley (1969) and Richard Benton et al. (1995). Features such as the difficulty of content words, average sentence length, standardised type:token ratios and the use of function words were used in different combinations to create seven methods to measure text difficulty. Teachers’ and students’ ratings of text difficulty, and students’ scores on reading comprehension tasks related to the texts were used as criteria to examine the validity of the readability methods. The findings revealed that indices of either vocabulary load or lexical density when used in combination with the number of function types in the text, produce statistical significance with the criterion measures. Further research is needed to confirm their validity for use in Māori –medium classroom settings. The Māori word lists developed for this project as the basis of the readability approaches have the potential for more widespread analyses of language proficiency measures for students in Māori-medium settings.