Made available in DSpace on 2016-04-28T18:23:24Z (GMT). No. of bitstreams: 1 LAEL - Maria Otilia Guimaraes Ninin.pdf: 2933716 bytes, checksum: 8358a23a1e0515c4559eee84d2fcdade (MD5) Previous issue date: 2006-10-26 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This research aims to discuss and characterize the Collaborative Research, considered as an activity system that offers an important context for the critical-collaborative construction of the knowledge in educational environments. In order to delineate the differences between collaborative research and collaborative actions in the research , I discussed the actions of the research, in order to bring about in the researchers, a reflection about their responsibility in relation to the researched contexts, considering that the collaborative research presumes the intervention of those involved in direction of their own transformation and, above all, to the researched situation. In this way, this work chose as a base for the research study, a continued professional development project in school, based on the relations between the researcher and professionals who work as director and pedagogical coordinator in state schools in the region of the Greater São Paulo. From the data collected in this context and with the discoursive analysis of the interactions, the collaborative research was discussed. As a foundation to the discussions, this study is based on the social-culturalhistorical activity theory (Vygotsky, 1934/2000, 1978/2000; Engeström, 1987, 1993, 1995, 1999, 2001, 2003, 2005; Daniels, 1993, 2001, 2004, 2005, 2006, and others), on the discussions about qualitative methodology of research (Bredo & Feinberg, 1982; Cole & Knowles, 1993; Demo, 2004), on the theory of pedagogical discourse (Bernstein, 1971, 1990, 1993, 2000), on the role of the language discussed by the systemic functional linguistics (Halliday, 1973, 1985.1994; Halliday & Hasan, 1989, and others), and, basically, on the discussions about collaboration (Magalhães, 1990, 1992, 1994, 196, 1998, 2000, 2002, 2003; Brookfield & Preskill, 1999; Bray et al., 2000, and others). Parts of reports or transcriptions about meetings between researcher and participants were selected and analyzed, chosen amongst a set of 55 meetings. The intention of analyzing the excerpts was to understand how the methodological actions that characterize the collaborative research process became themselves the organizers of the participants learning. In relation to the transformationpermanence of the activity systems, the results show that its elements, when were worked by means of the actions of the expansive learning net, were redressed or partially redressed for researcher and participants. The results also show that the relationship developed between researcher and participants were redressed and reorganized by the actions of the collaborative research / Esta pesquisa tem como objetivo principal discutir e caracterizar a Pesquisa Colaborativa, considerada como um sistema de atividade que propicia um importante contexto para a construção crítico-colaborativa do conhecimento, em ambientes educacionais. Em busca de delinear as diferenças entre pesquisa colaborativa e ações colaborativas na pesquisa , discuti as ações da pesquisa, a fim de desencadear nos pesquisadores uma reflexão sobre sua responsabilidade em relação aos contextos pesquisados, uma vez que a pesquisa colaborativa pressupõe a intervenção dos envolvidos em direção à sua própria transformação e, acima de tudo, da situação pesquisada. Nessa perspectiva, este trabalho escolheu como base para estudos de pesquisa um projeto de formação contínua pautado nas relações entre a pesquisadora e profissionais que exercem cargos de direção e coordenação pedagógica em escolas públicas na região da Grande São Paulo, e, a partir dos dados coletados nesse contexto e da análise das interações discursivas, discutiu a pesquisa colaborativa. Para fundamentar as discussões, o trabalho procurou subsídios nos pressupostos teórico-filosóficos da pesquisa sócio-histórico-cultural (Vygotsky, 1934/2000, 1978/2000; Engeström, 1987, 1993, 1995, 1999, 2001, 2003, 2005; Daniels, 1993, 2001, 2004, 2005, 2006, dentre outros), nas discussões sobre metodologia no paradigma qualitativo de pesquisa (Bredo & Feinberg, 1982; Cole & Knowles, 1993; Demo, 2004), na teoria do discurso pedagógico (Bernstein, 1971, 1990, 1993, 2000), no papel da linguagem como discutido pela lingüística sistêmico-funcional (Halliday, 1973, 1985.1994; Halliday & Hasan, 1989, dentre outros), e, fundamentalmente, nas discussões sobre colaboração (Magalhães, 1990, 1992, 1994, 196, 1998, 2000, 2002, 2003; Brookfield & Preskill, 1999; Bray et al., 2000, dentre outros). Para proceder às análises e discussões, foram selecionados recortes de relatos ou de áudio-gravações entre pesquisadora e pesquisandas, escolhidos dentre um conjunto de 55 encontros. A análise percorreu os excertos escolhidos com o propósito de entender como as ações metodológicas características de um processo colaborativo de pesquisar eram, também, organizadoras da aprendizagem dos participantes. Os resultados apontam que, em relação à transformação-permanência dos sistemas de atividade, seus elementos, trabalhados por meio das ações da rede de aprendizagem expansiva, foram, tanto pela pesquisadora, quanto pelas pesquisandas, ora ressignificados, ora parcialmente ressignificados. Apontam, ainda, que as relações desenvolvidas entre pesquisador e pesquisandos, a partir das ações da pesquisa colaborativa, foram ressignificadas e reorganizadas

Les concepts de topos, graphè, éthos et pathos contribuent activement à marquer la singularité d’une œuvre poétique, en l’occurrence, dans le cadre de notre thèse, celle de Louis Aragon (1897- 1982).Aragon fut l’inspirateur de compositeurs et d’interprètes comme « Kosma, Douai, Ferré, Ferrat, Brassens, Claveau, Montand, Trenet, Béart, Morelli, Martin, Ogeret ». Il déclare même qu’il aime qu’on lise ses poèmes mais préfère qu’on les chante. Outre l’importance de la dimension visuelle, la composante sonore et rythmique s’avère elle aussi capitale dans la perception et la réception de la poésie d’Aragon. C’est la raison pour laquelle nous avons opté dans notre thèse pour une étude analytique sémiostylistique et lexico-sémantique de l’œuvre poétique aragonienne.La sémiostylistique est « l’étude du fonctionnement sémiotique du style d’un texte ». En nous inspirant des travaux de G. Molinié, nous avons tenté d’analyser le « conglomérat lexico-syntactico-rhétorico-thématique » qui caractérise la poésie aragonienne en tentant d’ « apprécier la littérarité » des « segments occurrents, et [de] détecter leur spécificité ». Cela nous a permis de poser la question des conditions de la littérarité, de sa mesure, de sa valeur à réception et donc de l’effet de l’art à travers des procédés d’unité ou, au contraire, de subversion. / The concepts topos, graph, ethos and pathos actively contribute to highlight the singularity of a poetic work, such as that of Louis Aragon (1897-1982), the object of study of the present thesis. Aragon inspired many composers and singers such as "Kosma, Douai, Shoed, Ferrat, Brassens, Archstone, Montand, Trenet, Béart, Morelli, Martin, Ogeret".He even declares that he likes that his poems to be read but prefers are sung. Apart from the importance of the visual dimension, the sound and rhythmic components are also capital in the perception and the reception of the poetry of Aragon. This is why I opted in the present thesis for a semiostylistic and lexico-semantic analytical study of Aragon’s works of poetry. Semiostylistics is "the study of the semiotic operation of the style of a given text". Inspired by the works of G. Molinié, I tried to analyse the "conglomerate of lexico-syntactic, rhetorical and thematic characteristics" of Aragon’s poetry while trying "to appreciate the literariness" of the "occurring segments, and to detect their specificity" which enabled me to explore the conditions of the literary value, its measurement, its value at reception and, thus, the effect of art through the processes of unity or, on the contrary, subversion.

Cette thèse aborde la spécification et la réalisation de CELINE, outil de correction des erreurs basé sur une architecture multi-agents à deux niveaux : <BR> 1) Un système lourd, renfermant l'ensemble du savoir linguistique (multi-domaines par rapport à l'univers du discours), et générateur de systèmes individualisés. Les agents peuvent être considérés comme imparfaits ou partiellement inadaptés. Ils sont mis en concurrence par domaine d' expertise.<BR> 2) Un système léger implanté sur le site du rédacteur, système construit par apprentissage par le système central à partir des travaux de ce rédacteur.<BR> La spécification du système se construit tout au long des chapitres.<BR> La problématique de la correction des erreurs et la finalité de la conception d'un système de correction le plus automatique possible, avec des prises de décision à faible granularité reposant sur des critères multi-niveaux, nous entraînent vers un besoin de coopération justifiant une réalisation multi-agents.<BR> Une taxinomie des erreurs et des rappels sur l'analyse linguistique nous permet d'établir un début de structure du tableau noir du système. Nous consolidons nos choix par une comparaison du système attendu avec quelques prototypes du domaine. Nous examinons ensuite les comportements sociaux de deux agents chargés de définir l'un un modèle linguistique partiel suffisant du rédacteur et l'autre un sous-ensemble pertinent du système global. Nous découvrons alors le modèle de communication des agents et complétons notre structure de données par les marques de validité.<BR> La méthode des structures permet une quantification, incluse dans le tableau noir, de la correction des fautes d'accords.<BR> Après une approche des systèmes multi-agents, nous présentons une synthèse de l'architecture de CELINE et du fonctionnement des pilotes et de quelques agents.<BR> Un bilan rapide, précèdera en conclusion, une mise en situation du système proposé dans le cadre des industries de la langue et dans un environnement réseau du type Internet.

LE DISCOURS POLITIQUE RELATIF À L'AMENAGEMENT LINGUISTIQUE EN FRANCE (1997-2002) L'histoire de France est marquée depuis le XVIe siècle par l'uniformisation linguistique. La République a ouvert son ère par une Terreur politique qui s'est accompagnée de Terreur linguistique. Depuis, France et français sont intimement liés dans l'organisation comme dans les imaginaires politiques. Or, à un moment récent et bref de l'histoire de France, lors de la XIème législature (1997-2002), le débat a émergé quant à l'opportunité de reconnaitre une diversité linguistique de moins en moins importante sur le territoire national, les locuteurs des langues régionales disparaissant progressivement par un pur effet démographique. En effet, le débat sur la Charte européenne des langues régionales ou minoritaires (1999) puis sur le statut de la Corse (2001) a occupé la scène politique et médiatique française comme rarement les questions de statut des langues en France l'avaient fait. La multiplicité des lieux d'expression et des conditions de production et de réception des discours politiques a nécessité, pour aborder ce que les médias nomment « la classe politique » et que nous définissons comme une communauté discursive, la construction d'un corpus fortement hétérogène. Séances parlementaires à l'Assemblée nationale ou au Sénat, rapports, avis, projets ou propositions de loi, questions au gouvernement, mais également expression de la communauté discursive des hommes et des femmes politiques dans la presse écrite et audiovisuelle ont été réunis pour tenter de saisir le débat dans son ensemble. L'hétérogénéité constitutive du corpus a justifié un traitement différencié des sous corpus, en fonction de leur lieu de production et de leurs conditions de transmission : le corpus parlementaire, représentant plus de 250000 mots a fait l'objet d'un traitement automatique par Lexico3, ce qui a permis d'entrer dans le corpus. Le traitement lexicométrique de l'ensemble parlementaire et traitement manuel des corpus médiatiques ont été articulés de manière féconde : une analyse de discours à entrée lexicale a été possible grâce à la façon dont le traitement automatique a mis en valeur des phénomènes de catégorisation opérées par les locuteurs au moyen du lexique. L'approche lexico-sémantique a été complétée d'une cartographie des arguments en présence : la communauté discursive des hommes politiques dessine des imaginaires sociodiscursifs. Des idéologies concurrentes de ce qu'est la Nation et de son devenir s'opposent alors.

L'objectif de cette thèse est d'étudier le développement du système lexico-sémantique chez les enfants monolingues et bilingues. La question posée est la suivante : quand et comment les significations des mots commencent à être reliées entre elles et à s'intégrer dans un système sémantique interconnecté. Dans un premier temps, trois études ont été menées chez des enfants monolingues français. L'Etude 1, a pour but d'observer si les mots sont organisés selon des liens taxonomiques (e.g., cochon - cheval). L'Etude 2 explore si l'effet d'amorçage sémantique est sous-tendu par des mécanismes cognitifs, comme les processus d'activation automatique et contrôlé. Puis enfin, l'Etude 3 observe si les mots sont organisés en fonction de leur distance de similarité sémantique (e.g., vache - mouton versus vache - cerf). Dans un deuxième temps, deux études ont été conduites chez des enfants apprenant deux langues simultanément. L'Etude 4 vise à déterminer si les mots sont taxonomiquement liés dans chacune des langues. L'Etude 5 explore si les mots présentés dans une langue activent leurs représentations sémantiques dans l'autre langue et vice versa. Dans le but de répondre à ces questions, le traitement lexico-sémantique a été étudié en utilisant deux techniques : l'eye-tracking et les potentiels évoqués (PEs). Ces deux techniques enregistrent lors de la présentation des mots des réponses comportementales (Etude 3) et neuronales (Etude 1, 2, 4 et 5) de haute résolution temporelle. Les Etudes 1 et 2 montrent que chez les monolingues les mots sont liés taxonomiquement à l'âge de 18 et 24 mois. Durant le développement du langage, les deux processus d'activation automatique et contrôlé sont impliqués dans le traitement des mots (Etude 2). L'Etude 3 montre qu'à 24 mois, les mots sont organisés dans le système lexico-sémantique en développement selon la distance des similarités sémantiques. L'Etude 4 montre que chez les enfants bilingues, le traitement sémantique ne diffère pas selon les deux langues, mais la topographie des PEs varie selon la langue traitée. L'Etude 5 montre que les mots présentés dans une langue activent leurs représentations sémantiques dans la deuxième langue et vice versa. Toutefois, la topographie des PEs est modulée selon la direction de traduction. Ces résultats suggèrent que l'acquisition de deux langues, bien qu'elle soit très précoce, requière deux ressources neuronales bien distinctes, sous-tendant ainsi le traitement lexico-sémantique des langues dominante et non-dominante. / The present doctoral research explored the developing lexical-semantic system in monolingual and bilingual toddlers. The question of how and when word meanings are first related to each other and become integrated into an interconnected semantic system was investigated. Three studies were conducted with monolingual French learning children which aimed at exploring how words are organized, that is, according to taxonomic relationships (e.g., pig - horse) and to semantic similarity distances between words (e.g., cow - sheep versus cow - deer), and whether cognitive mechanisms, such as automatic activation and controlled processes, underlie priming effects. An additional two studies conducted with children learning two languages simultaneously, aimed at determining, first, whether taxonomically related word meanings, in each of the two languages, are processed in a similar manner. The second goal was to explore whether words presented in one language activate words in another language, and vice versa. In an attempt to answer these questions, lexical-semantic processing was explored by two techniques: eye-tracking and event-related potentials (ERPs) techniques. Both techniques provide high temporal resolution measures of word processing but differ in terms of responses. Eye-movement measurements (Study III) reflect looking preferences in response to spoken words and their time-course, whereas ERPs reflect implicit brain responses and their activity patterns (Study I, II, IV, and V). Study I and II revealed that words are taxonomically organized at 18 and 24-month-olds. Both automatic and controlled processes were shown to be involved in word processing during language development (Study II). Study III revealed that at 24-month-olds, categorical and feature overlap between items underpin the developing lexical-semantic system. That is, lexical-items in each semantic category are organized according to graded similarity distances. Productive vocabulary skills influenced word recognition and were related to underlying cognitive mechanisms. Study IV revealed no differences in terms of semantic processing in the bilinguals¿ two languages, but the ERP distribution across the scalp varied according to the language being processed. Study V showed that words presented in one language activate their semantic representations in the second language and the other way around. The distribution of the ERPs depended, however, on the direction of translation. The results suggest that even early dual language experience yields distinct neural resources underlying lexical-semantic processing in the dominant and non-dominant languages during language acquisition.

L'analyse sémantique de textes nécessite en préalable la construction d'objets relevant de la sémantique lexicale. Les vecteurs d'idées et les réseaux lexicaux semblent de bons candidats et constituent ensemble des structures complémentaires. Toutefois, faut-il encore être capable dans la pratique de les construire. Les vecteurs d'idées peuvent être calculés à partir de corpus de définitions de dictionnaires, de thésaurus ou encore de textes. Ils peuvent se décliner en des vecteurs conceptuels, des vecteurs anonymes ou des vecteurs lexicaux - chaque type présentant un équilibre différent entre précision, couverture et praticité. Quant aux réseaux lexicaux, ils peuvent être acquis efficacement via des jeux, et c'est précisément l'objet du projet JeuxDeMots. L'analyse sémantique peut être abordée par l'analyse thématique, et ainsi servir de moyen de calcul à des vecteurs d'idées (bouclage). Nous pouvons modéliser l'analyse comme un problème d'activation et de propagation. La multiplicité des critères pouvant intervenir dans une analyse sémantique, et la difficulté inhérente à définir une fonction de contrôle satisfaisante, nous amène à explorer l'usage de métaheuristiques bio-inspirées. Plus précisément, nous introduisons un modèle d'analyse par colonies de fourmis artificielles. A partir d'un texte, l'analyse vise a construire un graphe contenant les objets du texte (les mots), des objets identifiés comme pertinents (des syntagmes, des concepts) ainsi que des relations pondérées et typées entre ces objets.

[pt] A idiomaticidade das cores em vocábulos e expressões da língua portuguesa no Brasil investiga o processo de compreensão do significado não literal de um corpus onde em cada exemplo há um termo denominativo de cor. Busca-se neste estudo descrever e analisar esse corpus segundo seu comportamento léxico, sintático, semântico, pragmático e cultural , em situações reais de comunicação, com foco no ensino de Português como Segunda Língua (PL2) e/ou Português para Estrangeiros (PLE). A análise mostra que a cor é, antes de tudo, um fenômeno cultural, e sua percepção está ligada à sensibilidade do homem perante a realidade que o cerca. No contexto comunicativo o brasileiro frequentemente demonstra sua afetividade ao chamar alguém de neguinho(a), ou lhe oferece respeito e deferência ao estender o tapete vermelho ou se entrega à hipérbole ao classificá-lo de botafoguense roxo. Assim, a linguagem se apropria dos termos cromáticos em novas combinações lexicais para dizer mais que simplesmente o negro da noite, o vermelho do sangue, o roxo das violetas. O aprendiz de PL2/PLE percebe que não é possível compreender o significado dessas novas combinações utilizando somente o sentido literal das palavras que as compõem. O conhecimento de fenômenos linguísticos como a idiomaticidade e a composicionalidade, abordados nesta análise, permitirá ao professor de PL2/PLE ampliar a variedade de exercícios e estratégias em sala de aula, com o objetivo de ajudar os alunos na compreensão do significado dos exemplos apresentados e na adequação de seu uso em contextos reais de comunicação. / [en] The idiomaticity of the colors in words and expressions of Portuguese language in Brazil leads us to investigate the process of understanding the nonliteral meaning of a corpus, as for in each example there is a denominative term for a color. In this study we attempt to describe and analyze this corpus according to its lexical, syntactic, semantic, pragmatic and cultural conduct, in realistic situations of communication, focusing on the teaching of Portuguese as a Second Language (PL2) and/or Portuguese for Foreigners (PLE). The analysis demonstrates that color is, before anything else, a cultural phenomenon, and its perception is linked to man s sensitivity when fronting the reality surrounding him. Whithin the context of the communication the Brazilian frequently reveals his emotions, as when calling someone neguinho, a term of endearment for the word black; or when showing someone respect and deference, extending o tapete vermelho (the red carpet); or resorting to hyperbole in the classification of a soccer team fan botafoguense roxo (purple). Thus, language invades chromatic terms in new lexical combinations to say more than simply the blackness of night, the redness of blood, or the purple of the violets. The student of PL2/PLE perceives that it is not possible to understand the significance of these new combinations utilizing only the literal sense of the words forming their composition. The knowledge of linguistic phenomena such as their idiomaticity and compositiveness here analyzed, will afford the teacher of PL2/PLE the chance to increase the variety of exercises and strategies in the classroom, with the objective of helping students understand the meaning of the given examples and adapt their use to real communication contexts.

[pt] Transparência, no sentido de translúcido, é um conceito presente em uma enorme variedade de âmbitos e, recentemente, tem sido explorado no contexto de software. Um software transparente é aquele que informa sobre si mesmo, disponibilizando informação sobre o que faz, como e porque o faz. A abordagem da transparência no contexto de software se iniciou através da instanciação de um catálogo de requisitos não funcionais (RNFs), que inicialmente foi desenvolvido para transparência de processos. Durante a construção deste catálogo, notou-se que os RNFs dependência, rastreabilidade e detalhamento, presentes no catálogo, estão relacionados a problemas clássicos da Engenharia de Software. Estes problemas são, respectivamente, a definição de critérios para modularização do software, rastreabilidade entre seus artefatos e entendimento do seu código. Neste trabalho, propomos estratégias para lidar especificamente com estes três problemas, contribuindo também para construção de softwares mais transparentes, uma vez que sua solução implica na satisfação de qualidades do catálogo. Estas estratégias foram estruturadas através de um processo de construção de software. Esse processo tem como escopo software para Web, utilizando arquitetura MVC, desenvolvido através do paradigma de programação procedural. A ideia central do processo é a utilização de dois artefatos de requisitos, Léxico Ampliado da Linguagem (LAL) e cenários, durante todo processo de construção do software. O LAL é utilizado na definição dos módulos do software. Os cenários são refinados em diferentes níveis de abstração, de modo que permeiam todo o processo, desde os requisitos até serem integrados ao código fonte do software. Os rastros entre os diferentes níveis de abstração dos cenários são mantidos. As operacionalizações propostas no processo foram anexadas ao catálogo de transparência de software, possibilitando futura reutilização orientada a qualidades. O processo e suas operacionalizações foram avaliados através de estudos empíricos envolvendo diversos sujeitos. O resultado das avaliações sugere que o processo contribui positivamente para as qualidades de detalhamento, dependência e rastreabilidade, presentes no catálogo de transparência. / [en] Transparency is a concept that permeates different disciplines, and recently has been explored in the context of software engineering. Transparent software aims to make information about it transparent and also aims to provide transparency on its behavior. The study of transparency in the context of software started by instantiating a catalog of non-functional requirements (NFRs). During the construction of this catalog, we realized that the NFRs of traceability, dependability and detailing are related to classical problems in software engineering. These problems are, respectively, the definition of software modules, traceability between its artifacts and the understanding of its source code. In this work we propose strategies to deal specifically with these three problems. As these problems are also related to the transparency catalog, then if we contribute to solving these problems, we will also be contributing to software transparency. These strategies were structured through a software construction process that has Web software as its main target. The process uses MVC architecture and procedural oriented programming languages. The central idea of this process is to use two requirements artifacts, the Language Extended Lexicon (LEL) and scenarios, throughout the construction process. The LEL is used in the definition of the software modules. The scenarios are refined at different levels of abstraction so that they permeate the entire development process, from requirements documentation until their integration with the software source code. The traces between these scenarios are maintained. The strategies proposed in the process were attached to the software transparency catalog to allow for quality based reuse. The process and its operationalizations were evaluated through empirical studies involving different subjects. The results of the evaluations suggest that the process contributes positively to the quality of detailing, dependency and traceability, present in the transparency catalog.

[pt] Este trabalho constitui-se num estudo morfológico sobre a questão produtividade / produção de padrões lexicais em textos literários, com o objetivo de sustentar a hipótese de que tais obras apresentam condições de produção especiais em relação a determinados padrões morfológicos de formação. A base de investigação é a obra O Coronel e o Lobisomem, de José Candido de Carvalho, já conhecida por seu teor de inovação lexical. São eleitas para análise as sufixações com –ão e –oso e a prefixação com des-, formações morfológicas que se constituem em potenciais processos de inovação lexical. Sugere-se que a utilização de tais padrões na obra não é fortuita, já que parecem representar iconicamente o discurso hiperbólico do personagem principal. De início, apresenta-se uma visão geral da morfologia lexical, quanto a seu tratamento tradicional e em relação a contribuições mais recentes, focalizando-se especialmente a questão produtividade/produção. Procede-se depois à apreciação do tratamento dado pelos estudiosos do português sobre os formativos com –ão, -oso e des-. Em seguida, são analisadas as formações encontradas na obra O Coronel e o Lobisomem. Finalmente, são apresentadas as conclusões sobre o uso que o autor faz dos padrões de formação em sua obra, os quais refletem o discurso do personagem principal, com vistas a afirmar a relevância da análise das condições de produção, ao lado da questão da produtividade, para um melhor conhecimento do léxico do português. / [en] This work is a morphologic study on the issue of lexical productivity and production in literary texts. Its main goal is to support the hypothesis that literary works present special conditions for the use of morphological formation patterns.Our investigation is based on the novel The Colonel and the Werewolf by Jose Candido de Carvalho, a work which is known for its lexical innovations. The suffixes –ão and –oso, as well as the prefix des- were chosen for our analysis, as they are the preferred morphological elements for lexical innovation in Carvalhos s work. Our results suggest that the use of these morphological patterns in the work is not fortuitous; in fact, they seem to iconically represent the hyperbolic discourse of the main character of the novel. We initially present an overview of lexical morphology, which includes both traditional approaches and recent contributions; the main focus will be on the productivity/production issue. Then, we approach the description of the suffixes –ão and –oso and the prefix desin Portuguese. Next, the lexical constructions with these elements occurring in the book The Colonel and the Werewolf are analyzed. We conclude showing that the author uses morphological patterns to reinforce the discourse of the main character of the novel. This analysis shows the relevance of the study of production conditions, in addition to productivity, for a deeper understanding of the lexicon in Brazilian Portuguese.

[pt] Neste trabalho estudam-se os nomes com a estrutura NdeN que podem ser descritos como nomes compostos por justaposição. São observadas 1.500 seqüências de palavras, com o objetivo de contribuir na descrição formal do léxico do português do Brasil e de definir os critérios de identificação de um nome composto com essa estrutura. O critério geral está baseado no conceito da nãocomposicionalidade semântica. Os testes são feitos a partir das propriedades sintáticas e semânticas que há na relação entre os elementos que constituem o grupo nominal, mostrando as distinções entre um grupo nominal livre e um grupo nominal composto. Entre as propriedades, podem ser destacadas: o bloqueio distribucional, a inseparabilidade, a inserção lexical, o apagamento de N2, a substituição de N2 e as variações em gênero e número. Essa descrição mostra-se útil na medida em que um conjunto de regras e critérios de delimitação de unidades lexicais foi definido, constituindo uma base para a incorporação de novos itens ao léxico. Por fim, as propriedades das entradas incorporadas receberam uma representação formal, resultando na criação de um dicionário eletrônico utilizável em processos eletrônicos. / [en] This paper is a study of the NofN structure nouns, which may be described as compound nouns by juxtaposition. 1500 word sequences are observed, aiming at contributing to the formal description of the Brazilian Portuguese lexicon, and defining the identification criteria of a compound noun with such a structure. The general criterion is based on the concept of semantic non- compositionality. The tests are made from the syntactic and semantic properties existing in the relationship between the elements that constitute the nominal group, showing the differences between a free nominal group and a compound nominal group. Among such properties, the following can be pointed out: distributional blockage, inseparability, lexical insertion, N2 erasing, N2 substitution, and gender and number variations. Such description proves to be useful in the sense that a set of lexical units delimitation rules and criteria has been defined, constituting a basis for the incorporation of new items to the lexicon. Finally, the incorporated entries` properties received a formal representation, which resulted in the creation of an electronic dictionary that can be used in electronic processes. / [fr] On étudie dans ce travail les noms formés par la structure NdeN qui caractérise un nom composé par juxtaposition. On a observé 1.500 séquences de mots, pour définir les critères d`identification d`un nom composé par cette structure. Le critère general est fondé sur le concept de la non compositionalité sémantique. Les tests ont été faits à partir des propriétés syntaxiques et sémantiques qu`il y a dans la relation existante entre les éléments qui constituent le groupe nominal, en démontrant les distinctions entre un groupe nominal libre et un groupe nominal composé. Parmi ces propriétés, on peut remarquer le bloquement distributionnel, l`inséparabilité, l`insertion lexical, l`effacement de N2, le remplacement de N2 et les variations en genre et en nombre. En examinant ces propriétés, on a pu observer qu`il y a des irrégularités dans la formation de ce procès. Cette description devient utile dans la mesure que cette reconnaissance peut être formulée et qu`un ensemble de règles et de critères de délimitation d`unités lexicales peut être defini, et que, par là, on peut aboutir à l`incorporation de nouveaux items lexicaux. Au bout, en tenant compte de la possibilité de formalisation, les mots composés peuvent être processés automatiquement dans les diccionnaires électroniques.