No description available.

This dissertation is an application of the framework of Distributed Morphology to the morphosyntax of Chemehuevi, an endangered Southern Numic language of the Uto-Aztecan family. Following one of the central claims of DM, I argue that word formation in Chemehuevi happens in the syntax and provide evidence for this claim from the formation of lexical categories, as well as from the morphosyntax of the Chemehuevi causative verbs. I frame my discussion of lexical categories around the Root Hypothesis (Marantz 1997, Arad 2005), a notion that there are no underived nouns, verbs, or adjectives in the grammar, but roots that receive interpretation and assignment to a `part of speech' depending on their functional environment. I show that Chemehuevi nouns and verbs are formed when roots are incorporated into nominal or verbal functional heads, many of which are overtly represented in the language. I also demonstrate that there is no distinct class of adjectives in Chemehuevi, and that roots with adjectival meanings are derived into stative verbs or nominalizations, depending on their function.My discussion of predication in Chemehuevi centers around the previously unexplained distribution of the enclitic copula -uk, which under my analysis is viewed as an overt realization of a functional head Pred (based on Baker 2003), which is obligatory in the formation of nominal and adjectival, but not verbal predicates.Another major theme of the dissertation is the notion that word-formation from roots differs from word-formation from derived words, known as the Low vs. High Attachment Hypothesis (Marantz 2000, Travis 2000, etc.). This approach explains the differences between compositional and non-compositional word formation by the distance between the root and functional head(s) attached to it. On the basis of Chemehuevi causatives, I show that causative heads attached directly to the root derive words that exhibit morphophonological and semantic idiosyncrasies, such as allomorphy and availability of idiomatic meanings, while high attachment heads derive words that are fully compositional. This locality constraint on interpretation of roots is explained in terms of phase theory, and I present evidence from Chemehuevi showing that what constitutes a phase may be subject to parametric variation.Each chapter of the dissertation contains a section for non-linguistic audience where I provide a summary of the main points in non-theoretical terms and connect them to practical applications for the purposes of language learning and revitalization.

In this thesis we investigate two operational models of name-passing calculi: one based on coalgebra, and one based on enriched automata. We develop a semantic framework for modelling the open bisimulation in π-calculus, hyperbisimulation in Fusion calculus, and the first semantic interpretation of FOλ(nabla) logic. We consider a category theoretic model where both “variables” and “names”, usually viewed as separate notions, are particular cases of the more general notion of atoms. The key aspect of this model is to consider functors over the category of irreflexive and symmetric finite relations. The models previously proposed for the separate notions of “variables” and “names” embed faithfully in the new one, and initial algebra/final coalgebra constructions can be transferred from the formers to the latter. Moreover, the new model admits a definition of distinction-aware simultaneous substitutions. As a substantial application example, we give the first semantic interpretation of Miller-Tiu's FOλ(nabla) logic. FOλ(nabla) logic is designed for reasoning about operational semantics of languages with binding. On the operational level, a contribution of the thesis is to extend an automata-based model for a variety of name-passing calculi with their associated notion of equivalence. HD-automata, a syntax-independent operational model, has been successfully applied for modelling e.g. early and late bisimulation in π-calculus and hyperbisimulation in Fusion calculus. However, current HD-automata are not adequate for modelling open bisimulation because they can not handle distinction-preserving substitutions. We solve this technical problem by integrating the notion of distinction into the definition of named sets, the basic building blocks of HD-automata. Finally, we discuss the relationship between HD-automata with distinctions, and D-LTS.

The first part of this thesis addresses the following questions: where in the syntactic tree, and at what representational level is an NP Case-checked. To this end, it presents converging data from French, Welsh and Irish, which suggest (i) that Case-checking may be accomplished under a variety of functional projections (subject to parametric variation); and (ii) that Case positions are--at least partially--independent of the A/A$ sp prime$-distinction. It furthermore presents evidence from Irish and Welsh--VSO languages in which NPs typically raise to their Case position only at LF--that NPs are, under certain conditions, Case-checked at S-structure. / Chapter 2 investigates word order and cliticisation in Standard French and Quebec French interrogatives and proposes a typology of interrogatives. Chapter 3 and 4 account for complementizer variation, pre-verbal particles and agreement patterns in Welsh and Irish under a Case-theoretic approach. / The second part of this thesis concerns the conditions on the availability of structural accusative Case. A theory of structural Case is proposed according to which accusativity is a configurational rather than a lexical property--i.e., resulting from syntactic geometry and not from lexical feature specifications on verbs. To this end, a comparison between the syntactic mapping of stative and perfective predicates in Irish and English is undertaken.

Any community that wants to be part of technological progress has to ensure that the language(s) of that community has/have the necessary human language technology resources. Part of these resources are so-called "core technologies", including part-of-speech taggers. The first part-of-speech tagger for Afrikaans is developed in this research project. It is indicated that three resources (a tag set, a twig algorithm and annotated training data) are necessary for the development of such a part-of-speech tagger. Since none of these resources exist for Afrikaans, three objectives are formulated for this project, i.e. (a) to develop a linpsticdy accurate tag set for Afrikaans; (b) to deter- mine which algorithm is the most effective one to use; and (c) to find an effective method for generating annotated Afrikaans training data. To reach the first objective, a unique and language-specific tag set was developed for Afrikaans. The resulting tag set is relatively big and consists of 139 tags. The level of specificity of the tag set can easily be adjusted to make the tag set smaller and less specific. After the development of the tag set, research is done on different approaches to, and techniques that can be used in, the development of a part-of-speech tagger. The available algorithms are evaluated by means of prerequisites that were set and in doing so, the most effective algorithm for the purposes of this project, TnT, is identified. Bootstrapping is then used to generate training data with the help of the TnT algorithm. This process results in 20,000 correctly annotated words, and thus annotated training data, the hard resource which is necessary for the development of a part-of-speech tagger, is developed. The tagger that is trained with 20,000 words reaches an accuracy of 85.87% when evaluated. The tag set is then simplified to thirteen tags in order to determine the effect that the size of the tag set has on the accuracy of the tagger. The tagger is 93.69% accurate when using the diminished tag set. The main conclusion of this study is that training data of 20,000 words is not enough for the Afrikaans TnT tagger to compete with other state-of-the-art taggers. The tagger and the data that is developed in this project can be used to generate even more training data in order to develop an optimally accurate Afrikaans TnT tagger. Different techniques might also lead to better results; therefore other algorithms should be tested. / Thesis (M.A.)--North-West University, Potchefstroom Campus, 2005.

Dispersion of mass is a measure of the deviation of transportation of fluid in a reactor from ideal reactor behavior (perfect mixing or plug flow) caused by the combined effects of diffusion, convection and migration. Axial dispersion is always undesirable because it reduces the driving force of the reaction and therefore causes a lower level of conversion. On the other hand, transverse dispersion is often a desirable feature since good transverse mixing will reduce the transverse concentration and temperature gradients and hence improve the selectivity of a thermochemical reactor. Transverse dispersion of mass is of more importance in a three-dimensional flow-by electrochemical reactor than that in a thermochemical reactor because the potential drop is in the transverse direction and the reaction rate and selectivity are determined by the potential as well as concentration and temperature distributions. The transverse dispersion of mass is expected to have a more profound effect on the performance of a 3D electrochemical reactor due to the strong interaction among the concentration, temperature and potential distributions in the transverse direction. In the present work, the axial and transverse dispersion of mass were studied with a twodimensional dispersion model in two types of rectangular packed bed: i) randomly packed glass beads with the average bead diameter of 2 mm and a macroscopic bed porosity of 0.41; ii) a representation of a 3D flow-by electrode - consisting of a bed of carbon felt with the carbon fibre diameter of 20 urn and a macroscopic bed porosity of 0.95. A tracer stimulation-response system was set up and axial and transverse dispersion of In Blackfoot, DPs appear to take obligatory wide scope with respect to the universal quantifier while bare nouns take obligatory narrow scope with respect to the universal quantifier. I propose that the difference in scope-taking properties of Blackfoot nominals is a consequence of their syntactic position. I propose that over argument DPs are adjoined to the clause whereas bare nouns are base generated in an argument position. I suggest that the scope properties fall out from this distinction in the syntax. The Blackfoot universal quantifier, ohkan-, is a preverb. That is, ohkan- occurs as a part of the verb stem preceding the verb root itself. I propose that ohkan- is head of its own QP which takes the VP as its complement. I follow Sportiche (1998) in categorizing ohkan- as a stranded quantifier since it is base generated external to VP. Bare nouns, since they are generated within VP, are structurally inferior to ohkan-, since they are within its c-command domain. The adjoined DPs, however, are structurally superior to ohkan-, since they are adjoined to the clause. I propose that the structural superiority of DPs translates to their obligatory wide scope. Conversely, the structural inferiority of bare nouns translates to their obligatory narrow scope. Blackfoot is a relatively understudied Algonquian language spoken in Southern Alberta and Northern Montana. The Blackfoot data presented in this work come primarily from my own work with two Blackfoot speakers. Both of my language consultants hail from Southern Alberta speak and the Blood dialect of Blackfoot.

Lithuanian language is quite in an early stage of language processing. And therefore has a high demand on automated tools like taggers, parsers, word sense disambiguators etc. During the last 10 years only a few researchers were attempting to create a parser for Lithuanian language. However none of them are used in practices nowadays. The process of designing and implementing rule based parser for Lithuanian language is presented in this paper. Rules and constraints of the formal grammar follow the principles of Dependency Grammar. Necessary language recourses were made up at the Computational Centre of Linguistics. Research area analysis and overview of the most popular methodologies is presented here as well. Syntax parser of the Lithuanian language was evaluated against the Gold Standard and gave 80,2 % of accuracy of in recognizing parts of the sentence.

Magistro studijų baigiamojo darbo „Broniaus Kalinausko asmenybė ir mokslinė veikla“ tikslas – pateikti B. Kalinausko biografijos fragmentus bei atskleisti, koks kalbininko indėlis į lietuvių kalbos frazeologiją, sintaksę, kalbos kultūrą. Taip pat apžvelgti B. Kalinausko metodinius darbus, skirtus tiek mokytojams, tiek mokiniams. Darbe stengiamasi ne tik pabrėžti B. Kalinausko skelbtas idėjas, jo požiūrį į tam tikrus kalbos dalykus, bet ir palyginti iškeltas mintis su vėlesnių kalbininkų darbais, atskleisti, koks pamatas padėtas tolimesniems lietuvių kalbos tyrinėjimams. Atkreipiamas dėmesys į tai, kaip B.Kalinausko pateikta lietuvių kalbos frazeologijos klasifikacija pakito iki šių dienų, kaip kitos iškeltos idėjos įsitvirtino lietuvių kalbotyroje. Analizuojami sintaksės, metodikos darbai, straipsniai, susiję su kalbos praktika. Darbo pabaigoje pateikiami kalbininką pažinojusių žmonių prisiminimai, kurie leidžia geriau pažinti B. Kalinausko asmenybę. / The main purpose of a final master‘s degree thesis „The personality and scientific activities of Bronius Kalinauskas“ is to introduce to B. Kalinauskas biographical snippets and to disclose any linguist‘s contribution to the Lithuanian language phraseology, syntax and language culture. It is also an overview of B. Kalinauskas methodological work, for both teachers and students. This thesis attempts to emphasize not only B. Kalinauskas published ideas, his approach to certain language issues, but also in relation to the ideas raised in the subsequent work of linguists, to reveal what lay the foundation for further research in Lithuanian language. The attention is drawn to it how B. Kalinauskas the Lithuanian language phraseology clasification changed to these days. There are also analysed syntax, methodology works, articles related to the language practice. The end of a final master‘s degree thesis provides memories of his acquaintances, which allows a better understanding of B. Kalinauskas personality.

This thesis proposes a generalized probabilistic approach to modelling document collections along the combined axes of both semantics and syntax. Probabilistic topic (or semantic) models view documents as random mixtures of unobserved latent topics which are themselves represented as probabilistic distributions over words. They have grown immensely in popularity since the introduction of the original topic model, Latent Dirichlet Allocation (LDA), in 2004, and have seen successes in computational linguistics, bioinformatics, political science, and many other fields. Furthermore, the modular nature of topic models allows them to be extended and adapted to specific tasks with relative ease. Despite the recorded successes, however, there remains a gap in combining axes of information from different sources and in developing models that are as useful as possible for specific applications, particularly in Natural Language Processing (NLP). The main contributions of this thesis are two-fold. First, we present generalized probabilistic models (both parametric and nonparametric) that are semantically and syntactically coherent and contain many simpler probabilistic models as special cases. Our models are consistent along both axes of word information in that an LDA-like component sorts words that are semantically related into distinct topics and a Hidden Markov Model (HMM)-like component determines the syntactic parts-of-speech of words so that we can group words that are both semantically and syntactically affiliated in an unsupervised manner, leading to such groups as verbs about health care and nouns about sports. Second, we apply our generalized probabilistic models to two NLP tasks. Specifically, we present new approaches to automatic text summarization and unsupervised part-of-speech (POS) tagging using our models and report results commensurate with the state-of-the-art in these two sub-fields. Our successes demonstrate the general applicability of our modelling techniques to important areas in computational linguistics and NLP.

No description available.