Computer Sciences / In order to respond to increasing demand for natural language interfaces---and provide meaningful insight into user query intent---fast, scalable lexical semantic models with flexible representations are needed. Human concept organization is a rich phenomenon that has yet to be accounted for by a single coherent psychological framework: Concept generalization is captured by a mixture of prototype and exemplar models, and local taxonomic information is available through multiple overlapping organizational systems. Previous work in computational linguistics on extracting lexical semantic information from unannotated corpora does not provide adequate representational flexibility and hence fails to capture the full extent of human conceptual knowledge. In this thesis I outline a family of probabilistic models capable of capturing important aspects of the rich organizational structure found in human language that can predict contextual variation, selectional preference and feature-saliency norms to a much higher degree of accuracy than previous approaches. These models account for cross-cutting structure of concept organization---i.e. selective attention, or the notion that humans make use of different categorization systems for different kinds of generalization tasks---and can be applied to Web-scale corpora. Using these models, natural language systems will be able to infer a more comprehensive semantic relations, which in turn may yield improved systems for question answering, text classification, machine translation, and information retrieval. / text

Analyses of English verb meaning often rely on quasi-aspectual operators embedded in event structures to explain shared properties across classes. These operators scope over temporally basic meaning elements that make up the idiosyncratic semantic core of complex verbs. While the inventory of operators – or semantic primes – differ from proposal to proposal, they are generally presented as a closed class that includes at least CAUSE and BECOME, and their presence and location in event structures account for several alternation and ambiguity phenomena. In this study, I investigate a number verbs whose decompositions would include only operator(s) and event structure frames under most current decompositional lexical theories; in particular, the periphrastic causatives (cause, make, etc) and the verbs of becoming (become, get, etc). I account for differences in the selectional behavior of these verbs by positing incorporated meaning components beyond the purely aspectual or event structural. Based in part on regularities among corpus collocations, I propose additional meaning distinctions among these verbs along the parameters of causal patient complicity, sentiment, and register. / text

No description available.

Os documentos oficiais e a concretização das propostas no livro didático analisado abordam a ambiguidade lexical como algo a ser evitado e corrigido através de estratégias de desambiguação, ignorando a importância da natureza polissêmica da língua como recurso expressivo na comunicação cotidiana e como instrumento fundamental do humor, especialmente as variantes de humor popular e infantil. O presente trabalho busca abordar a ambiguidade lexical, polissemia e homonímia, no gênero tirinha cômica, texto multimodal e humorístico, visando observar a eficácia da proposta de leitura e produção como exercício de contextualização, levando os alunos a observarem, explicarem e produzirem texto, utilizando-se dos diversos significados possíveis das palavras em uso, em situação concreta de enunciação. A proposta se dá dentro do conceito de multimodalidade, trabalhando a ambiguidade lexical e o letramento visual, objetivando que o aluno não apenas observe e compreenda as tirinhas apresentadas, mas também que produza a sua própria, compondo o texto com os elementos verbal e visual. As atividades foram divididas em dez passos: pesquisa e compartilhamento de HQs, atividade diagnóstica, análise da ambiguidade lexical em tirinha cômica no livro didático, jogo digital Quem ri seus males espanta Piadas, consulta às diferentes acepções no dicionário, escrita do glossário de unidades lexicais polissêmicas ou homônimas, elaboração do roteiro da tirinha, autoavaliação, produção final de uma tirinha cômica e apreciação dos trabalhos das turmas. / The analised official papers and textbook proposals approach lexical ambiguity as something to be avoided and correct through desambiguation strategies, ignoring the importance of the polissemic nature of language as an expressive resource of the language in daily communication and as a fundamental tool of comedy and humor, especially popular and childrens jokes and puns. This paper aims to approach lexical ambiguity, polissemy and homonyms, within the comic strip genre, which is a multimodal and comic text genre, with the objetive of observing how effective this proposal of reading and writing can be to help students comprehend, learn and explain lexical ambiguity, by using a word in multiple possible meanings in the comic strip, in concrete enunciation. This proposal uses multimodality, by combining lexical ambiguity and visual litteracy, aiming to enable the student not only to observe and understand the comic strips presented, but also to produce their own by using verbal and visual components. The sequence is divided in ten steps: research and sharing comics, diagnostic evaluation, lexical ambiguity in the textbook, digital game Quem ri seus males espanta Piadas, searching for different meanings of words in the dictionary, writing a glossary of polissemic unities, production of a scratch comic, self evaluation, final production and appreciation of the classes works.

This thesis describes a novel methodology, grounded in the distributional semantic paradigm, for building context sensitive models of word meaning, affording an empirical exploration of the relationship between words and concepts. Anchored in theoretical linguistic insight regarding the contextually specified nature of lexical semantics, the work presented here explores a range of techniques for the selection of subspaces of word co-occurrence dimensions based on a statistical analysis of input terms as observed within large-scale textual corpora. The relationships between word-vectors that emerge in the projected subspaces can be analysed in terms of a mapping between their geometric features and their semantic properties. The power of this modelling technique is its ability to generate ad hoc semantic relationships in response to an extemporaneous linguistic or conceptual situation. The product of this approach is a generalisable computational linguistic methodology, capable of taking input in various forms, including word groupings and sentential context, and dynamically generating output from a broad base model of word co-occurrence data. To demonstrate the versatility of the method, this thesis will present competitive empirical results on a range of established natural language tasks including word similarity and relatedness rating, metaphor and metonymy detection, and analogy completion. A range of techniques will be applied in order to explore the ways in which different aspects of projected geometries can be mapped to different semantic relationships, allowing for the discovery of a range of lexical and conceptual properties for any given input and providing a basis for an empirical exploration of distinctions between the semantic phenomena under analysis. The case made here is that the flexibility of these models and their ability to extend output to evaluations of unattested linguistic relationships constitutes the groundwork for a method for the extrapolation of dynamic conceptual relationships from large-scale textual corpora. This method is presented as a complement and a counterpoint to established distributional methods for generating lexically productive word-vectors. Where contemporary vector space models of distributional semantics have almost universally involved either the factorisation of co-occurrence matrices or the incremental learning of abstract representations using neural networks, the approach described in this thesis preserves the connection between the individual dimensions of word-vectors and statistics pertaining to observations in a textual corpus. The hypothesis tested here is that the maintenance of actual, interpretable information about underlying linguistic data allows for the contextual selection of non-normalised subspaces with more nuanced geometric features. In addition to presenting competitive results for various computational linguistic targets, the thesis will suggest that the transparency of its representations indicates scope for the application of this model to various real-world problems where an interpretable relationship between data and output is highly desirable. This, finally, demonstrates a way towards the productive application of the theory and philosophy of language to computational linguistic practice.

La communication en général, et les capacités linguistiques en particulier, constituent de formidables obstacles pour beaucoup d'enfants avec TSA, qui ont du mal à faire connaître leurs idées, sentiments et intentions à autrui. La sémantique lexicale est fondamentale dans le choix des bons mots et la compréhension du monde autour de soi. Dans la recherche sur le Trouble du spectre de l’autisme (TSA), les capacités langagières sont communément mesurées sur la base de la performance sur des tâches qui mesurent les capacités en sémantique lexicale—les connaissances sur la signification des mots. Or, les connaissances scientifiques sur comment les capacités sémantiques lexicales sont liées aux autres aspects du langage dans le TSA sont pauvres. / Communications in general and linguistic abilities in particular, constitute formidable obstacles for many children with ASD, who struggle with making their ideas, feelings and intentions known to others. Lexical semantics is fundamental to choosing the right words, and understanding the surrounding world. In research on ASD, language abilities are very widely measured on the basis of performance on tasks testing lexical semantic abilities—knowledge of the meaning of words. However, very little is known about how lexical semantic abilities are related to other aspects of language in ASD.

The present study explores the concepts behind Japanese and Swedish dimensional adjectives. The focus is on examining which similarities and differences in the conceptualisation across the two languages exist, if any at all. In order to see how concepts underlying dimensional adjectives are represented in the speaker's minds, data was collected mainly from two word-association tests. The results show that dimensional adjectives are conceptualised and represented differently by speakers of these two languages. The most remarkable difference resulting from the word-association tests is that Japanese participants associate dimensional adjectives mostly with nouns that denote entities the prominent extension of which is aptly described by the stimulus dimensional adjective (e.g. 'long' is associated with 'river'). In Swedish, however, participants associate dimensional adjectives with both adjectives and nouns, and the association patterns and their underlying conceptualisations are thus more diverse (e.g. 'high' is associated with 'building', and 'long' is associated with 'narrow').

In the Bantu language Kinyarwanda, the morpheme –ish can be used to mark both causation and the instrumental applicative. This report pro- poses an explanation for this causative-instrumental syncretism, arguing that both causation and the introduction of an instrument are—at their core—two outgrowths of the same semantic notion. Fitting with other morphological causatives in Bantu, the causative use of –ish patterns as a lexical causative marker. The analysis presented here captures the lex- ical nature of the causative use of the morpheme by arguing that the new causal link is added sub-lexically, situating Kinyarwanda into a cross- linguistic typology of morphological causatives. / text

Orientador: Bento Carlos Dias da Silva / Banca: Clotilde de Almeida Azevedo Murakawa / Banca: Patrícia Tosqui Lucks / Resumo: A cidade de Ibitinga destaca-se nacionalmente no ramo dos bordados, com sua economia voltada quase que exclusivamente à produção dessas peças. Tendo esse contexto como pano de fundo, esta dissertação discute todo o processo linguístico e linguístico-computacional de construção de ontoléxicos - constructos formais cuja natureza léxico-conceitual possibilita o desenvolvimento de investigações teóricas (estudos lexicogramaticais) e aplicadas (construção de recursos lexicográficos e para o processamento computacional de informação textual disponível na Web). Em particular, constrói um ontoléxico exploratório que entrelaça conceitos e itens lexicais do domínio léxico-conceitual da Indústria do Bordado de Ibitinga. O embasamento teórico-metodológico assenta-se nos estudos de semântica lexical (wordnets) e de frames (framenets), pura e computacional, de processamento automático de língua natural e de ontologias. / Abstract: Ibitinga, which is a Brazilian town nationally known for its outstanding position in the embroidery business, has its economy almost totally based on the production of a broad selection of embroideries. With this embroidery industry in the backdrop, this master thesis describes the whole linguistic and computational-linguistic process of construction of ontolexicons - formal constructs whose lexical-conceptual nature allows for both theoretical (lexical-grammar construction) and applied (lexicographical and ontolexical resource construction) research. In particular, the study focuses on the design and implementation of a toy ontolexicon for the Ibitinga Embroydery Industry lexical-conceptual domain. Its theoretical foundations have drawn heavily on lexical semantics (wordnets), frame semantics (framenets), natural language processing, and ontologies. / Mestre

Os documentos oficiais e a concretização das propostas no livro didático analisado abordam a ambiguidade lexical como algo a ser evitado e corrigido através de estratégias de desambiguação, ignorando a importância da natureza polissêmica da língua como recurso expressivo na comunicação cotidiana e como instrumento fundamental do humor, especialmente as variantes de humor popular e infantil. O presente trabalho busca abordar a ambiguidade lexical, polissemia e homonímia, no gênero tirinha cômica, texto multimodal e humorístico, visando observar a eficácia da proposta de leitura e produção como exercício de contextualização, levando os alunos a observarem, explicarem e produzirem texto, utilizando-se dos diversos significados possíveis das palavras em uso, em situação concreta de enunciação. A proposta se dá dentro do conceito de multimodalidade, trabalhando a ambiguidade lexical e o letramento visual, objetivando que o aluno não apenas observe e compreenda as tirinhas apresentadas, mas também que produza a sua própria, compondo o texto com os elementos verbal e visual. As atividades foram divididas em dez passos: pesquisa e compartilhamento de HQs, atividade diagnóstica, análise da ambiguidade lexical em tirinha cômica no livro didático, jogo digital Quem ri seus males espanta Piadas, consulta às diferentes acepções no dicionário, escrita do glossário de unidades lexicais polissêmicas ou homônimas, elaboração do roteiro da tirinha, autoavaliação, produção final de uma tirinha cômica e apreciação dos trabalhos das turmas. / The analised official papers and textbook proposals approach lexical ambiguity as something to be avoided and correct through desambiguation strategies, ignoring the importance of the polissemic nature of language as an expressive resource of the language in daily communication and as a fundamental tool of comedy and humor, especially popular and childrens jokes and puns. This paper aims to approach lexical ambiguity, polissemy and homonyms, within the comic strip genre, which is a multimodal and comic text genre, with the objetive of observing how effective this proposal of reading and writing can be to help students comprehend, learn and explain lexical ambiguity, by using a word in multiple possible meanings in the comic strip, in concrete enunciation. This proposal uses multimodality, by combining lexical ambiguity and visual litteracy, aiming to enable the student not only to observe and understand the comic strips presented, but also to produce their own by using verbal and visual components. The sequence is divided in ten steps: research and sharing comics, diagnostic evaluation, lexical ambiguity in the textbook, digital game Quem ri seus males espanta Piadas, searching for different meanings of words in the dictionary, writing a glossary of polissemic unities, production of a scratch comic, self evaluation, final production and appreciation of the classes works.