Ordering of representational level adverbial modifiers in spoken Brazilian Portuguese

Supported by the Functional Discourse Grammar theoretical model, as proposed by Hengeveld (2005), this paper aims to show that the order of modifiers of the Representational Level in spoken Brazilian Portuguese is determined by scope relations according to the layers of property, state-of-affairs and propositional content. This kind of distribution indicates that, far from being free-ordered as suggested by traditional grammarians, modifiers have a preferred position determined by semantic relations that may be only changed for pragmatic and structural reasons.

Children’s and adults’ processing of syntactically ambiguous sentences during reading

While there has been a fair amount of research investigating children’s syntactic processing during spoken language comprehension, and a wealth of research examining adults’ syntactic processing during reading, as yet very little research has focused on syntactic processing during text reading in children. In two experiments, children and adults read sentences containing a temporary syntactic ambiguity while their eye movements were monitored. In Experiment 1, participants read sentences such as, ‘The boy poked the elephant with the long stick/trunk from outside the cage’ in which the attachment of a prepositional phrase was manipulated. In Experiment 2, participants read sentences such as, ‘I think I’ll wear the new skirt I bought tomorrow/yesterday. It’s really nice’ in which the attachment of an adverbial phrase was manipulated. Results showed that adults and children exhibited similar processing preferences, but that children were delayed relative to adults in their detection of initial syntactic misanalysis. It is concluded that children and adults have the same sentence-parsing mechanism in place, but that it operates with a slightly different time course. In addition, the data support the hypothesis that the visual processing system develops at a different rate than the linguistic processing system in children.

Structure informationnelle et constructions du kabyle : Etude de trois types de phrase dans le cadre de la grammaire constructionnelle

Information structure and Kabyle constructions Three sentence types in the Construction Grammar framework The study examines three Kabyle sentence types and their variants. These sentence types have been chosen because they code the same state of affairs but have different syntactic structures. The sentence types are Dislocated sentence, Cleft sentence, and Canonical sentence. I argue first that a proper description of these sentence types should include information structure and, second, that a description which takes into account information structure is possible in the Construction Grammar framework. The study thus constitutes a testing ground for Construction Grammar for its applicability to a less known language. It constitutes a testing ground notably because the differentiation between the three types of sentences cannot be done without information structure categories and, consequently, these categories must be integrated also in the grammatical description. The information structure analysis is based on the model outlined by Knud Lambrecht. In that model, information structure is considered as a component of sentence grammar that assures the pragmatically correct sentence forms. The work starts by an examination of the three sentence types and the analyses that have been done in André Martinet s functional grammar framework. This introduces the sentence types chosen as the object of study and discusses the difficulties related to their analysis. After a presentation of the state of the art, including earlier and more recent models, the principles and notions of Construction Grammar and of Lambrecht s model are introduced and explicated. The information structure analysis is presented in three chapters, each treating one of the three sentence types. The analyses are based on spoken language data and elicitation. Prosody is included in the study when a syntactic structure seems to code two different focus structures. In such cases, it is pertinent to investigate whether these are coded by prosody. The final chapter presents the constructions that have been established and the problems encountered in analysing them. It also discusses the impact of the study on the theories used and on the theory of syntax in general.

Adolescent writing development: Improving succinctness using the extended noun phrase

This study examined an aspect of adolescent writing development, specifically whether teaching secondary school students to use strategies to enhance succinctness in their essays changed the grammatical sophistication of their sentences. A quasi-experimental intervention was used to compare changes in syntactic complexity and lexical density between one-draft and polished essays. No link was demonstrated between the intervention and the changes. A thematic analysis of teacher interviews explored links between changes to student texts and teaching approaches. The study has implications for making syntactic complexity an explicit goal of student drafting.

Mas e embora: uma descrição morfossintática e semântico-pragmática

Esta dissertação é o resultado de dois anos de estudo e de pesquisa dedicados à formação, à estrutura e ao funcionamento da língua portuguesa. A inconformidade com o tratamento um tanto superficial que os compêndios gramaticais prescritivos, sobretudo os escolares, têm dado à classificação das conjunções coordenativas adversativas e das subordinativas adverbiais concessivas e, por conseguinte, das orações em que elas se inserem, foi a motivação e é o que justifica a escolha do tema. Essa maneira de abordar a questão tem misturado descrições sintáticas com semânticas, sem, muitas vezes, levar em consideração a aplicabilidade e a pertinência de uma ou de outra oração no discurso. Dessa forma, esta dissertação se propõe a descrever os comportamentos morfossintáticos e semântico- -pragmáticos dos períodos formados com as conjunções mas e embora à luz da semântica enunciativa, e a averiguar até que ponto a vagueza de abordagem desse tema tem refletido na formação daqueles que têm a língua portuguesa como língua materna no Brasil. Para tanto, o trabalho apresenta a seguinte estrutura: num primeiro momento, serão considerados a metodologia usada para a feitura do trabalho e os pressupostos teóricos que melhor fundamentam-no. A teoria está dividida em três momentos, a saber, o enfoque dado pela tradição gramatical, o enfoque dado pela linguística à noção de coordenação e de subordinação, e o enfoque dado pela semântica enunciativa às conjunções mas e embora. Num segundo momento, serão explanadas as análises feitas dos corpora presentes no texto: um para corroborar as ideias presentes em um dos capítulos teóricos (corpus de textos de opinião); e outro para averiguar se a superficialidade da abordagem do tema ao longo do tempo tem afetado na formação de leitores e escritores proficientes em língua portuguesa (corpus de transcrição). A análise do segundo corpus o corpus de transcrição, em que pessoas de diferentes níveis de estudo foram testadas no que tange ao conhecimento acerca da transformação de períodos compostos por coordenação adversativa em subordinação adverbial concessiva e vice-versa; além de se descobrir se não iniciariam uma frase com a conjunção adversativa mas tem capítulo à parte e revelará o que não é tão difícil de pressupor: 67,4% das frases transcritas revelam o desconhecimento que as pessoas têm acerca da diferença morfossintática e semântico-pragmática de períodos formados pelas conjunções adversativa e concessiva, mas e embora, respectivamente

Phrase-based statistical language generation using graphical models and active learning

Most previous work on trainable language generation has focused on two paradigms: (a) using a statistical model to rank a set of generated utterances, or (b) using statistics to inform the generation decision process. Both approaches rely on the existence of a handcrafted generator, which limits their scalability to new domains. This paper presents BAGEL, a statistical language generator which uses dynamic Bayesian networks to learn from semantically-aligned data produced by 42 untrained annotators. A human evaluation shows that BAGEL can generate natural and informative utterances from unseen inputs in the information presentation domain. Additionally, generation performance on sparse datasets is improved significantly by using certainty-based active learning, yielding ratings close to the human gold standard with a fraction of the data. © 2010 Association for Computational Linguistics.