The aim of this research has been to describe and analyse Swedish linguistic and cultural influence, on identification, and idiomatic awareness as well as on the Spanish used by Hispanic Americans who have lived more than 12 years in Sweden. The analysis has been carried out within the field of sociolinguistics, specifically in the area of language contact, drawing on methodological aspects of ethno linguistics, sociology of language and cognitive linguistics. The research was divided into two phases. In the first phase, we describe and analyse Corpus 1, consisting of 20 interviews of first generation Hispanic Americans political refugees. In the second phase, we describe, compare and analyse Corpus 2, which consists of 105 questionnaires distributed to first and second generation Hispanic Americans in Sweden. The results have shown that in Sweden Hispanic Americans have gradually generated a variety of Spanish which clearly shows features of Swedish culture and language and with an aspect of standardisation due to the fact that in the intercommunication between language speakers of varieties of Spanish certain variations have slowly been eliminated over the course of time. The sociolinguistic situation of the informants has been observed as well as the nature of the immigration and identification and idiomatic awareness in relation to their integration or lack of integration into Swedish society. The use of 38 words with a high level of synonymy was also investigated in order to establish standardisation of these representative words among the varieties of Spanish, spoken in Sweden. Furthermore we have investigated clichés, lexical, semantic-pragmatic and syntactic loans from Swedish as well as Swedish cultural influence on use tense among the informants.

The overall aim of the present study is to investigate syntactic variation in certain Present-day English noun phrase types including the quantifiers all, whole, both and half (e.g. a half hour vs. half an hour). More specific research questions concerns the overall frequency distribution of the variants, how they are distributed across regions and media and what linguistic factors influence the choice of variant. The study is based on corpus material comprising three newspapers from 1995 (The Independent, The New York Times and The Sydney Morning Herald) and two spoken corpora (the dialogue component of the BNC and the Longman Spoken American Corpus). The book presents a number of previously not discussed issues with respect to all, whole, both and half. The study of distribution shows that one form often predominated greatly over the other(s) and that there were several cases of regional variation. A number of linguistic factors further seem to be involved for each of the variables analysed, such as the syntactic function of the noun phrase and the presence of certain elements in the NP or its near co-text. For each of the variables, all factors were ranked according to their strength of correlation with particular variants. The study also discusses a possible grammaticalisation process concerning NPs with half and the possibility of all sometimes having another function than expressing totality: to express large quantity. The whole idea of grammatical synonymy has been questioned by some scholars, but the conclusion drawn in the present study is that there are variables that are at least very close to each other in meaning, and that a number of linguistic and non-linguistic factors influence our choices of variant. A great deal of the information obtained was too detailed to be useful for pedagogical purposes, but in several cases the results could clearly be used to improve school and reference grammars.

The aim of the present study is to examine how the development of syntactic complexity and accuracy in spoken Spanish L2 is related to two parameters, motivation and aptitude, in which individual differences are often claimed to correlate with successful second language learning (Dörnyei and Skehan 2003). The study shows how an understanding of the relationship between progression in second language proficiency and motivation is gained through the study of longitudinal data and the assessment of motivation in discourse. The development of syntactic complexity and accuracy in oral performance has been chosen as main indicators of proficiency level. Correlations have been computed between proficiency levels, development coefficients, quantitative outcomes of aptitude tests and quantified assessment of motivation. The study is based on longitudinal data consisting of recorded interviews and conversations in Spanish in Swedish upper secondary school. The results from the progression analysis show that the whole group develops syntactic complexity, and that individual learners concentrate on either syntactic complexity or accuracy in their development. Discourse analytical methods have been applied to the in-depth interviews in order to identify motivational types expressed by the learner and the strength of each type. These have been represented in terms of values on a self-determination scale, the extreme points of which are 'intrinsic' vs. 'extrinsic' motivation. An aptitude test (Lat03, Meara 2005) has also been administered to the participants and analysed . Correlations (Spearman) have been further calculated between the relevant parameters. The results indicate that self-determined, intrinsic motivation correlates with progression. Analytic aptitude, however, does not predict progression to the same extent as intrinsic motivation seems to do at this level.

Current representation schemes for automatic text classification treat documents as syntactically unstructured collections of words (Bag-of-Words) or `concepts' (Bag-of-Concepts). Past attempts to encode syntactic structure have treated part-of-speech information as another word-like feature, but have been shown to be less effective than non-structural approaches. We propose a new representation scheme using Holographic Reduced Representations (HRRs) as a technique to encode both semantic and syntactic structure, though in very different ways. This method is unique in the literature in that it encodes the structure across all features of the document vector while preserving text semantics. Our method does not increase the dimensionality of the document vectors, allowing for efficient computation and storage. We present the results of various Support Vector Machine classification experiments that demonstrate the superiority of this method over Bag-of-Concepts representations and improvement over Bag-of-Words in certain classification contexts.

Current representation schemes for automatic text classification treat documents as syntactically unstructured collections of words (Bag-of-Words) or `concepts' (Bag-of-Concepts). Past attempts to encode syntactic structure have treated part-of-speech information as another word-like feature, but have been shown to be less effective than non-structural approaches. We propose a new representation scheme using Holographic Reduced Representations (HRRs) as a technique to encode both semantic and syntactic structure, though in very different ways. This method is unique in the literature in that it encodes the structure across all features of the document vector while preserving text semantics. Our method does not increase the dimensionality of the document vectors, allowing for efficient computation and storage. We present the results of various Support Vector Machine classification experiments that demonstrate the superiority of this method over Bag-of-Concepts representations and improvement over Bag-of-Words in certain classification contexts.

In der Arbeit wird die Entwicklung einer Herstellungstechnologie für Hohlkugeln (HK) und Schalen mit ihrem anschließenden Einsatz als Füllgut in einem syntaktischen Schaum mit Polymermatrix dargestellt. Die Technologie basiert auf einem Übergangsprozess der Erstarrung einer Zinnschmelze, in dem die Gestaltsgebung von HK durch die Wirkung eines Gasmediums stattfindet. Eine entsprechende Pilot-Anlage mit der Steuerung der Druck- oder Temperaturwerte wurde aufgebaut und unter verschiedenen technologischen Bedingungen getestet. Es wurde ein theoretisches Druck- und Temperaturmodell, die als Steuerfunktionen genutzt werden können, vorgeschlagen. Während der Entwicklungsstufe wurden die Konstruktion und die Simulation der relevanten physikalischen Prozesse durchgeführt und mithilfe von Nebenversuchen verifiziert. Zur qualitativen und quantitativen Bewertung der ermittelten Ergebnisse ist die statistische Versuchsplanung herangezogen worden. Anschließend sind die Erzeugnisse metallographisch und röntgenographisch untersucht worden.

Limited information exists regarding the acquisition of syntax and morphology in young Spanish-English bilingual language learners. A method to measure short-term language learning is through structural priming; an auditory model of the target structure is presented, which influences a subject’s subsequent production. The purpose of this thesis was to develop and pilot priming tasks in both English and Spanish to analyze the language production of typically developing bilingual elementary school students. The morphosyntactic structures targeted in the structural priming task included the third person singular and past tense in English, as well as direct object clitics and imperfect tense in Spanish. The study included three participants, aged 4;7, 6;7, and 10;11. Results revealed that bilingual elementary students with varied language exposure are influenced and learn from morphosyntactic priming. / text

本篇研究目的在於藉由分析句法直接性(syntactic directness)、語意修飾(semantic modification)以及內容 (content)來探討漢語母親規範語中的控制行為(control acts)，也就是指令(directives)和禁止(prohibitions)。語料來自一名以漢語為母語的母親與兒童之間的日常對話。研究結果顯示，指令隨著兒童年紀增長而遞減，而禁止卻隨著兒童年紀增長而遞增。句法直接性分析結果發現，不論指令或是禁止的情況下，漢語母親常使用祈使句(imperative)的句式。語意修飾方面結果發現，漢語母親主要是採用未修飾(bald)以及縮小(minimization)。內容方面顯示，指令常使用在從事能力活動(competent action)，而禁止常使用在得體行為(appropriate behavior)和自立(caretaking)。 / The purpose of this study is to investigate Mandarin maternal control acts including the directives and prohibitions in maternal regulatory language by analyzing syntactic directness, semantic modification, and content. The data collected were natural conversations of one Mandarin-speaking mother-child dyad. The data of maternal regulatory language was analyzed when the child’s ages were 2;1, 2;7, 3;1, and 3;7. The results show that the frequency of directives decreases with the child’s age, but the frequency of prohibitions increases. In addition, the preferred sentence type is imperative in both directives and prohibitions. The child’s cognitive development and the culture factors which would determine the style of regulatory language used by the Mandarin mother are discussed. Furthermore, the results of semantic modification reveal that bald and minimization are two dominant modifications. The mother’s adoption of bald and minimization may be influenced by her power status and politeness. Two major kinds of semantic combinations are also discovered in this study. Finally, the results of contents show that competent action often occurs in directives. As for prohibitions, appropriate behavior and caretaking are most related. Our finding would shed some light on the Mandarin maternal regulatory language.

Le principal enjeu de notre recherche est d’aboutir à un modèle intégratif fonctionnel pour l’analyse des verbes d’affect en français et en arabe. Nous avons choisi d’étudier quatre V_affect : deux verbes d’émotion (étonner et énerver en français et leurs équivalents [ʔadhaʃa], [ʕaɣḍaba] en arabe]) et deux verbes de sentiment (admirer et envier et leurs équivalents [ʔaʕʒaba] et [ħasada]) appartenant aux champs sémantiques de la surprise, la colère, l’admiration et la jalousie. Plus concrètement, l’analyse se situe :– au niveau sémantique et syntaxique : les dimensions sémantiques véhiculées par les collocatifs verbaux comme étonner tellement, énerver prodigieusement, en français, et [ʔaʕʒaba ʔiʕʒāban kabīran] (admirer admiration grand) , [ɣaḍaba ɣaḍabaan ʃadīdan] (énerver colère sévère), en arabe, sont systématiquement reliées à la syntaxe (les constructions grammaticales récurrentes) (Hoey, 2005).– au niveau syntaxique et discursif : les emplois actifs, passifs, et pronominaux des V_affect sont étudiés dans la perspective des dynamiques informationnelles au sein de la phrase (Van Valin et LaPolla, 1997).D’un point de vue méthodologique, l’étude s’appuie sur une démarche quantitative et qualitative de la combinatoire verbale et privilégie la démarche contrastive. Elle est fondée sur le corpus journalistique français de la base de données EmoBase (projet Emolex 100 M de mots) et du corpus journalistique ArabiCorpus (137 M de mots).La thèse contribue ainsi à l’étude des valeurs sémantiques, du comportement syntaxique et discursif de la combinatoire des V_affect, en arabe et en français, ce qui permet de mieux structurer le champ lexical des affects par rapport à ce que proposent les études existantes en lexicologie. Les principaux résultats de l’étude peuvent être appliqués en didactique des langues, en traductologie et en traitement automatisé du lexique des affects dans les deux langues comparées. / The paramount stake of this research is to achieve an integrative functional model for the analysis of affective verbs in French and Arabic. I have chosen four affective verbs: two verbs of emotion (to astonish and to rage in French and their equivalent in Arabic) and two verbs of sentiment (to admire and to envy in French and their equivalent [ʔadhaʃa], [ʔaɣḍaba] in Arabic) they belong to semantic dictions of Surprise, Anger, Admiration, and Jealousy. More concretely, the analysis is shaped:- On the semantic and syntactic level: the semantic dimensions carried by verbal collocations such as to extremely astonish, to rage prodigiously in French, and [ʔaʕʒaba ʔiʕʒāban kabīran] (admire admiration big)*, [ɣaḍaba ɣaḍabaan ʃadīdan] (to rage rage extreme), and in Arabic are systematically linked to syntax (the recurrent grammatical constructions) (Hoey 2005).- On the syntactic and discursive level: the usage of passive, active and reflexive forms of affective verbs are dealt with from the perspective of informational dynamics in the sentence. (Van Valin et LaPolla 1997).From a methodological point of view, the study is based on the quantitative and qualitative approach of the verbal combination and favours the contrastive one. It is founded on the French journalistic corpus of Emobase Database (Emolex project 100 M of words) and the journalistic corpus Arabicorpus) (137 M of words).Furthermore, the thesis participates in the studies of semantic values, the syntactic and the discursive behavior of affective verbs’combinations, in Arabic and in French, which will enable to better structure the diction of emotions in relation to what is proposed by current studies in lexicography. The main results of the study can be applied in language teaching, translation, and automated processing of emotions' lexicon in the two compared languages.

The goal of this thesis is to advance the use of recurrent neural network language models (RNNLMs) for large vocabulary continuous speech recognition (LVCSR). RNNLMs are currently state-of-the-art and shown to consistently reduce the word error rates (WERs) of LVCSR tasks when compared to other language models. In this thesis we propose various advances to RNNLMs. The advances are: improved learning procedures for RNNLMs, enhancing the context, and adaptation of RNNLMs. We learned better parameters by a novel pre-training approach and enhanced the context using prosody and syntactic features. We present a pre-training method for RNNLMs, in which the output weights of a feed-forward neural network language model (NNLM) are shared with the RNNLM. This is accomplished by first fine-tuning the weights of the NNLM, which are then used to initialise the output weights of an RNNLM with the same number of hidden units. To investigate the effectiveness of the proposed pre-training method, we have carried out text-based experiments on the Penn Treebank Wall Street Journal data, and ASR experiments on the TED lectures data. Across the experiments, we observe small but significant improvements in perplexity (PPL) and ASR WER. Next, we present unsupervised adaptation of RNNLMs. We adapted the RNNLMs to a target domain (topic or genre or television programme (show)) at test time using ASR transcripts from first pass recognition. We investigated two approaches to adapt the RNNLMs. In the first approach the forward propagating hidden activations are scaled - learning hidden unit contributions (LHUC). In the second approach we adapt all parameters of RNNLM.We evaluated the adapted RNNLMs by showing the WERs on multi genre broadcast speech data. We observe small (on an average 0.1% absolute) but significant improvements in WER compared to a strong unadapted RNNLM model. Finally, we present the context-enhancement of RNNLMs using prosody and syntactic features. The prosody features were computed from the acoustics of the context words and the syntactic features were from the surface form of the words in the context. We trained the RNNLMs with word duration, pause duration, final phone duration, syllable duration, syllable F0, part-of-speech tag and Combinatory Categorial Grammar (CCG) supertag features. The proposed context-enhanced RNNLMs were evaluated by reporting PPL and WER on two speech recognition tasks, Switchboard and TED lectures. We observed substantial improvements in PPL (5% to 15% relative) and small but significant improvements in WER (0.1% to 0.5% absolute).