Previous studies have examined either the input factors predicting language proficiency in bilingual children or the relationship between bilingualism and metalinguistic awareness. This thesis takes a novel approach exploring the two areas simultaneously. A study was conducted to investigate, first, the input factors that may cause variation in bilingual language proficiency and, secondly, the effects of differing levels of bilingualism on metalinguistic awareness. The participants were 38 French-English bilingual children aged six to eight, of middle to high socio-economic status, attending an international school in France. Data on the children's language experiences and family background were collected through questionnaires given to parents and children. Language proficiency was measured using the standardised French and English versions of the Peabody Picture Vocabulary Test. Metalinguistic awareness was assessed through seven metalinguistic tasks each given in both languages. The findings are discussed in relation to Bialystok's (1986a) analysis and control framework and Cummins' (1976) threshold hypothesis. The results indicated a strong association between language exposure estimates and language proficiency measures for each language. Furthermore, the child's stronger language was shown to be a reliable predictor of variables related to language use, including the language used with peers and the language the child finds easier to speak. The results for metalinguistic awareness were generally consistent with Bialystok's and Cummins' predictions. High level balanced bilinguals outperformed dominant bilinguals on high control tasks and on certain analysis tasks, but only when the child's best score, sometimes coming from the weaker language, was considered. A strong relationship was found between the language proficiency measures and the analysis tasks. Likewise, children scoring above the median on each of the Peabody tests generally outperformed those scoring below on analysis tasks. Overall, the results indicate that proficiency in each language, as well as degree of bilingualism, impact on metalinguistic awareness.

The thesis contends that the fundamental problem affecting the status and condition of minority languages is the lack of a sound economic base, due, historically, to a decline in the speech-communities' access to resources and the power to exploit them. In contrast, majority languages are supported by relatively vibrant economies, with knowledge of the languages concerned a necessary requirement for participation: the use of language and the viability of the economic base are reciprocal and mutually reinforcing. The intrusion of the language of an economically strong community into another community's work practices is the initial breach which commences the process of language shift; the extent to which a language is used in the world of work is an indicator of the strength of economic base. Following a description of the historical and social factors shaping the current circumstances of Gaelic, and an outline of the thesis's economic premise, a report is given on the results of a survey carried out to discover the extent of the use of Gaelic in the world of work, in particular the number of posts, and the fields of work, for which a knowledge of the language was deemed desirable or essential. To gain perspective on the Gaelic data, a comparative study is made of other minority languages, to explore the relationship between language condition and use in the work domain, and to discover any particular practices and policies utilised elsewhere which might be applicable to Gaelic. A review of relevant aspects of sociolinguistic theory is presented, with particular reference to the Gaelic situation: language use within society; language shift, decline, death and revival and restoration; the association of language with identity and nationalism, and the connection between language, economics, and language planning. The political context affecting Gaelic is analysed, to take cognisance of the constraints and opportunities moulding the present and future condition of the language.

<i>Part 1</i>: The distinction between a social rule and convergence of behaviour does not depict the internal point of view regarding behaviour in traditional rural social groups. Nor does it depict how people actually behave. It is ideological, contingent on the modern era and the emancipation of the individual. It is a matter of the theorist's presuming a logic, a form of life, that stipulates what may count as a reason for action. <i>Part 2</i>: Linguistic jurisprudence, as ordinary language philosophy that it is, does not expect, regarding the conceptual, to find a reality independent of society and the happenstance of society's language. In a positivist fashion, it merely describes social reality. The distinctions it comes up with (and so the one Part 1 was about) are factual instead, part of the logic/ form of life of modern Western society. - Yet social reality is complex and many sided, not a coherent theory. Existing assumptions/distinctions in social reality/ language are contradictory and interminable. - Linguistic jurisprudence is not concerned with all assumptions that may exist as possibilities, only with the typical ones that form the network of assumptions, which communication presupposes. - There are many languages not one. Communication exists no more, even if Oxonian armchair philosophy keeps taking it for granted. - Communicating is not <i>presupposing</i> rules. It is <i>changing</i> them, <i>adjusting</i> them to the people we encounter. We are both the same and different, there are both many and one languages. It is all a matter of what we choose to see. The positive reality of any given aggregate will entail an infinite number of communities/ societies, and corresponding languages, along with their negations. Uncommitted description cannot take place (is interminable) without the adoption by the theorist of a logic, which cannot be found in the external reality to be described. <i>Part 3</i>: Language is not a matter of uncommitted observation of social practice, but rules in our minds. The theorist is not reporting social groups' languages, but examining his own.

The primary objective of this dissertation is to provide a syntactic analysis of morphology in Estate Tamil, a dialect of Tamil spoken in tea estates, particularly in the Haputale area, in Sri Lanka. Tamil is significant for the theory of morphology not only because it is rich in a number of inflectional and derivational suffixes, but also because it exemplifies the co-occurrence of these two types of suffixes in the same lexical element, e.g. in participial and gerundive nouns. Further, these nouns have verbal morphology inside the nominal suffixes. Thus, the data from Tamil challenge the hypotheses that assume a distinction between inflection and derivation. Consequently, hypothesizing that all bound morphemes with selectional properties are functional heads, I explore morphosyntactic properties of deverbal nominals. In chapter 1, I claim that Estate Tamil is a distinctive dialect, contrary to the assumption that it is the same as Indian Tamil. Chapter 2 discusses some data and the questions that they raise for the theory of morphology. I argue that the criteria suggested for determining the distinction between inflection vs derivation and lexical rules vs syntactic rules are not adequate. Consequently, suggestion is made that all bound morphs with selectional properties can be analysed syntactically. Chapter 3 distinguishes lexical and functional categories and describe some theoretical assumptions in which morphology in Tamil is analysed in this dissertation. In Chapters 4 to 6, I analyse verbal morphology because the deverbal nominals contain verbal suffixes. Chapter 7 provides an account of morphology in deverbal nominals incorporating the verbal analysis given in the previous chapters. Finally, Chapter 8 describes the implications and consequences of the morphosyntactic analysis of Estate Tamil.

This thesis presents the development and implementation of a set of rules to generate intonational specifications for unrestricted text. The theoretical assumptions which motivate this work are outlined, and the performance of the rules is discussed with reference to various test corpora and formal evaluation experiments. The development of our rules is seen as a cycle involving the implementation of theoretical ideas about intonation in a text-to-speech conversion system, the testing of that implementation against some relevant body of data, and the refinement of the theory on the basis of the results. The first chapter introduces the problem of intonation in text-to-speech conversion, discusses previous practical and theoretical approaches to the problem, and sets out the general approach which is followed in subsequent chapters. We restrict the scope of our rules to generating <i>acceptable neutral</i> intonation, an approximation to <i>broad focus</i> (Ladd 1980), and we present a rule-development strategy based on the idea of a <i>default specification</i>) which can be successively refined) and on the principle of making maximum use of all the information available from text. The second chapter presents a framework for deriving an intonational specification in terms of <i>accents</i> and <i>boundaries</i> from a crude syntactic representation of any text sentence. This framework involves three stages: the division of text into intonational domains of various hierarchic levels; the assignment of accents to lexical items on the basis of stress information and grammatical class; and the modification of these accents and boundaries in accordance with phonological principles of prominence and rhythm. Chapter 3 discusses the problem of evaluating synthetic intonation, introduces an original evaluation procedure, and presents two formal evaluations of the output of the rules described in Chapter 2. Further sections present our attempts to improve our treatment of the three major causes of errors in the evaluated output: prepositional phrases, non-words or <i>anomalies</i> (e.g. numbers, dates and abbreviations), and anaphora of various kinds. The final chapter presents a summary of the main points of Chapters 1-3. We draw various conclusions regarding the nature of intonation, the development of text-to-speech conversion systems, and the generation of intonation in such systems.

The thesis investigates the link between memory and the construction of representations of individuals by studying referential change. The issue of how properties of individuals are represented as belonging to one and the same individual in memory is considered, and the processing of simple descriptions of pairs of individuals is considered in relation to this issue. Previous work has shown that referentially predictable descriptions of individuals impose no cost when attention switched between individuals. This investigation shows that when texts are referentially unpredictable, switching reference does cost time and that the two individuals are treated symmetrically. Recall analysis supports this and suggests that subjects focus on one of the individuals more than the other. This is explained by assuming that subjects encode simple surface characteristics which allow them to infer the pattern of attribute binding described by a text. When this mapping from surface order to semantics is disrupted by the introduction of unpredictability, subjects respond by treating the individuals asymmetrically which restores the mapping. It is proposed that this use of surface information is a general feature of language processing. Work on parallel function in pronoun comprehension is a closely related issue and some work had already addressed the issue of surface information effects. Therefore an investigation of parallelism in pronoun comprehension was carried out which revealed effects of surface information along with effects of grammatical parallelism which interacted with animacy and sentence structure. The investigation of switching reference also revealed word length effects which were interpreted as speech-based memory effects. Further investigations of these effects showed that they were not interfered with by articulatory suppression which contradicts interpretations of similar effects reported in the immediate serial list recall literature.

Artificial Intelligence techniques are increasingly being used to enhance the area of Computer-Aided Instruction. This thesis is concerned with the area of Computer-Aided Language Learning, a subset of Computer-Aided Instruction, and demonstrates how various Artificial Intelligence techniques can be incorporated into a language system to produce an intelligent educational tool. In this thesis, the focus is on the use of English articles, which is a subtle area of the English language with which even advanced students of English have difficulty. This thesis describes <i>Artcheck</i>, an intelligent Computer-Aided Language Learning (ICALL) system which detects, analyses and responds to English article usage errors. This system has three main features: it has knowledge of the article usage domain; it dynamically creates a <i>model</i> of the student; and it <i>adapts</i> to the individual student. The system's <i>knowledge</i> of the domain consists of a set of article usage rules which reflect standard teaching practice. The information necessary to apply the rules is extracted at the natural language processing stage, and includes structural and contextual information. The system <i>models</i> the state of the student's knowledge at all times, in order to give informative explanations to the student about any errors which are made. It is able to generate mal-rules which account for consistent errors made by the student, using <i>version spaces</i> and the <i>candidate elimination algorithm</i>. The student model can be described as <i>dynamic</i> because the generation of mal-rules can create new parts of the student model, in response to student behaviour, which are not pre-determined by the system designer. The system <i>responds</i> to individual students by giving explanations of errors which are tailored to the student's level of ability and preferred learning style.

This study is concerned with the investigation of the phenomenon of lexical cognateness and the transferability of French-English cognates from French (L2) to English (L3) by Burundian university students of English whose L1 is Kirundi, a Bantu language. The study examines cognate lexis according to the criteria of etymological origin as well as form, sound, and meaning correspondence and contrasts. It also investigates how 'false cognates' come about through semantic change and suggests the usefulness of examining French-English cognates in terms of 'cross-linguistic sense relations' such as synonymy, hyponymy, homonymy, and polysemy. These various features have been adopted as exploratory and classificatory devices in order to provide a linguistic framework for a systematic description and classification of French-English cognates. The investigation concludes that (I) apart from a few exceptions described as 'accidental cognates', French-English cognates are etymologically and historically related, (ii) false cognates have resulted from various forms of semantic change in the course of time such as the extension or narrowing of the original meaning, change in denotative meaning, and shift from denotative to figurative or emotive meaning, and (iii) a good number of French-English cognates can be described as cross-linguistic synonyms and hyponyms. The experimental part of the study comprises four tasks (sentence completion, acceptability judgement, multiple choice, and word association) all designed to assess the transferability of cognates from French to English. One hundred and twenty-six subjects selected from four different years of study in the Department of English language and literature of the University of Burundi took part in the experiments. The findings of the study support our three research hypotheses.

This thesis examines the occurrence and nature of modificational elements in phatic endphases. Phatic talk is an instance of interpersonal communication, it mainly has an ice-breaking function and is confined to the marginal phases of fact-to-face interaction. The collected speech material is contrastive in different respects. English production data are compared with interlanguage English and German data are compared with interlanguage German data. Phatic endphases are opposed to endphases which are more referential in nature. This approach makes it possible to draw up a cross-linguistic as well as an interlanguage phatic frame for different interpersonal configurations. A distinction is made between two types of encounter: interactions in which there is no social distance, i.e. between friends, and interactions with social distance, i.e. between, for instance, a shopkeeper and a client. An analysis of the data at all these levels reveals linguistic and socio-pragmatic deviations from a native speaker-like frame, which are explained at a cognitive level in terms of the learners' declarative and procedural knowledge. The learner data arising from the Discourse-Completion Task are supplemented with an extensive collection of introspection data. They make it possible to explain deviations in the learner's declarative and procedural knowledge not only at a cognitive, but also at an affective and social level.

No description available.