A criação de recursos linguístico-computacionais de base, como é o caso dos léxicos computacionais, é um dos focos da área de Processamento de Línguas Naturais (PLN). Porém, a maioria dos recursos léxicos computacionais existentes é específica da língua inglesa. Dentre os recursos já desenvolvidos para a língua inglesa, tem-se a VerbNet, que é um léxico com informações semânticas e sintáticas dos verbos do inglês, independente de domínio, construído com base nas classes verbais de Levin, além de possuir mapeamentos para a WordNet de Princeton (WordNet). Considerando que há poucos estudos computacionais sobre as classes de Levin, que é a base da VerbNet, para línguas diferentes do inglês, e dada a carência de um léxico para o português nos moldes da VerbNet do inglês, este trabalho teve como objetivo a criação de um recurso léxico para o português do Brasil (chamado VerbNet.Br), semelhante à VerbNet. A construção manual destes recursos geralmente é inviável devido ao tempo gasto e aos erros inseridos pelo autor humano. Portanto, há um grande esforço na área para a criação destes recursos apoiada por técnicas computacionais. Uma técnica reconhecida e bastante usada é o uso de aprendizado de máquina em córpus para extrair informação linguística. A outra é o uso de recursos já existentes para outras línguas, em geral o inglês, visando à construção de um novo recurso alinhado, aproveitando-se de atributos multilíngues/cross-linguísticos (cross-linguistic) (como é o caso da classificação verbal de Levin). O método proposto neste mestrado para a construção da VerbNet.Br é genérico, porque pode ser utilizado para a construção de recursos semelhantes para outras línguas, além do português do Brasil. Além disso, futuramente, será possível estender este recurso via criação de subclasses de conceitos. O método para criação da VerbNet.Br é fundamentado em quatro etapas: três automáticas e uma manual. Porém, também foram realizados experimentos sem o uso da etapa manual, constatando-se, com isso, que ela pode ser descartada sem afetar a precisão e abrangência dos resultados. A avaliação do recurso criado foi realizada de forma intrínseca qualitativa e quantitativa. A avaliação qualitativa consistiu: (a) da análise manual de algumas classes da VerbNet, criando um gold standard para o português do Brasil; (b) da comparação do gold standard criado com os resultados da VerbNet.Br, obtendo resultados promissores, por volta de 60% de f-measure; e (c) da comparação dos resultados da VerbNet.Br com resultados de agrupamento de verbos, concluindo que ambos os métodos apresentam resultados similares. A avaliação quantitativa considerou a taxa de aceitação dos membros das classes da VerbNet.Br, apresentando resultados na faixa de 90% de aceitação dos membros em cada classe. Uma das contribuições deste mestrado é a primeira versão da VerbNet.Br, que precisa de validação linguística, mas que já contém informação para ser utilizada em tarefas de PLN, com precisão e abrangência de 44% e 92,89%, respectivamente / Building computational-linguistic base resources, like computational lexical resources (CLR), is one of the goals of Natural Language Processing (NLP). However, most computational lexicons are specific to English. One of the resources already developed for English is the VerbNet, a lexicon with domain-independent semantic and syntactic information of English verbs. It is based on Levin\'s verb classification, with mappings to Princeton\'s WordNet (WordNet). Since only a few computational studies for languages other than English have been made about Levin\'s classification, and given the lack of a Portuguese CLR similar to VerbNet, the goal of this research was to create a CLR for Brazilian Portuguese (called VerbNet.Br). The manual building of these resources is usually unfeasible because it is time consuming and it can include many human-made errors. Therefore, great efforts have been made to build such resources with the aid of computational techniques. One of these techniques is machine learning, a widely known and used method for extracting linguistic information from corpora. Another one is the use of pre-existing resources for other languages, most commonly English, to support the building of new aligned resources, taking advantage of some multilingual/cross-linguistic features (like the ones in Levin\'s verb classification). The method proposed here for the creation of VerbNet.Br is generic, therefore it may be used to build similar resources for languages other than Brazilian Portuguese. Moreover, the proposed method also allows for a future extension of the resource via subclasses of concepts. The VerbNet.Br has a four-step method: three automatic and one manual. However, experiments were also carried out without the manual step, which can be discarded without affecting precision and recall. The evaluation of the resource was intrinsic, both qualitative and quantitative. The qualitative evaluation consisted in: (a) manual analysis of some VerbNet classes, resulting in a Brazilian Portuguese gold standard; (b) comparison of this gold standard with the VerbNet.Br results, presenting promising results (almost 60% of f-measure); and (c), comparison of the VerbNet.Br results to verb clustering results, showing that both methods achieved similar results. The quantitative evaluation considered the acceptance rate of candidate members of VerbNet.Br, showing results around 90% of acceptance. One of the contributions of this research is to present the first version of VerbNet.Br. Although it still requires linguistic validation, it already provides information to be used in NLP tasks, with precision and recall of 44% and 92.89%, respectively

La circonscription législative de Vire est une circonscription complexe de par la diversité des territoires qui la compose et donc par la diversité de son électorat. C’est une circonscription historiquement de droite, pourtant, entre 1958 et 2012, elle choisit par deux fois un député de gauche. Cette recherche vise donc à cerner et comprendre le comportement électoral des électeurs de la circonscription de Vire, ainsi que les facteurs qui déterminent leur choix une fois dans l’isoloir. Ces derniers sont multiples et ne pèsent pas tous de la même façon dans le choix du vote. Le contexte politique national, le contexte économique, l’électeur lui-même ou encore le candidat, tous ces facteurs y participent.Cette recherche se base avant tout sur des données quantitatives : résultats électoraux lors des élections législatives depuis 1958, mais aussi résultats électoraux lors des élections présidentielles. Il faut aussi étudier l’électorat par le biais des listes électorales, et des données publiées par l’INSEE. La composition de l’électorat de chaque commune, ainsi que les résultats électoraux associés sont donc répertoriés afin de permettre des comparaisons dans le temps, et entre communes. Ces relevés précis permettent également d’établir des cartes et de rendre l’évolution politique et sociologique de la circonscription plus lisible. Une fois mises en parallèle, ces données dressent le portrait de la circonscription de Vire à chaque scrutin législatif. Il convient également d’y ajouter ce que nous appelons la réalité du territoire. Nous nous reposons alors sur la mémoire vive ainsi que sur la presse locale pour rendre compte du contexte et des campagnes électorales.L’ensemble de ces facteurs : CSP, contextes etc., jouent sur le choix des électeurs de la circonscription étudiée. Il est pourtant un facteur qui caractérise la circonscription : la personnalité du candidat lui-même joue un un rôle dans le choix de l’électeur. Il aime connaître le candidat pour lequel il vote. La longévité d’Olivier Stirn en est une preuve, tout comme celle des députés suivants. C’est ainsi que la circonscription de Vire choisit par deux fois Alain Tourret, son seul député de gauche entre 1958 et 2012. / The french legislative division of Vire, in Normandy, is a complex division. It’s composed of different territories with specific electors. Historically, it’s a conservative division. However, the division chose a « left » (labour) deputy, twice, between 1958 and 1952.The goal of our research is to identify and explain the political behaviour of its electors, and reasons why they choose a candidate and not another. All reasons have a different importance between two electors. For example, the national political context, the economic context, or the elector and candidate personalities have different importance if the elector is a minor or a farmer.Quantitative datas are our first source. We use legislative elections results in the division since 1958, and presidential elections results too. We also study electors themselves by means of electoral lists and INSEE sources. Electors identities, and results of each common, written in tables and maps allow us to compare them. Thanks to this, we can draw a political and sociological portrait, during the 5th Republic, of the Vire division, and look at its evolutions. Our second source is less quantitative. We use local press et tracts to study legislative campaign. We use memories too. It shows how electors live in this territory.All these factors have a different importance, however, there is one which seems to be characteristic of Vire division. It appears that the elector likes to know the candidate he has chosen. Its personality is one of the most important factor. It explains Olivier Stirn’s carrier, and other deputies after him. We think it’s one of the reason why the Vire division elected Alain Tourret, twice, its only left (socialist) deputy.

This study focuses on the predicate argument structure (PAS) of a sub-class of verbs in Xitsonga - verbs of change of possession: give, contribute, future having, providing, obtaining and verbs of exchange. It is shown that these verbs select various theta roles to form their PAS in the different alternations allowed in this language. The effects of the applicative {-el-} and causative {-is-} verbal affixes on the PAS of such verbs are also considered. The study confirms the fact that the ordering of objects in ditransitive verbs is determined by an interplay of syntactic and semantic factors. Ambiguity arises in the case of two animate objects. In this case the object with a definite reading will appear adjacent to the verb. / African Languages / M. A. (Arican Languages)

This study focuses on the predicate argument structure (PAS) of a sub-class of verbs in Xitsonga - verbs of change of possession: give, contribute, future having, providing, obtaining and verbs of exchange. It is shown that these verbs select various theta roles to form their PAS in the different alternations allowed in this language. The effects of the applicative {-el-} and causative {-is-} verbal affixes on the PAS of such verbs are also considered. The study confirms the fact that the ordering of objects in ditransitive verbs is determined by an interplay of syntactic and semantic factors. Ambiguity arises in the case of two animate objects. In this case the object with a definite reading will appear adjacent to the verb. / African Languages / M. A. (Arican Languages)

Verbal Valency in a Cross-Linguistic Perspective Jana Šindlerová Abstract In the thesis, we look upon differences in argument structure of verbs considering the Czech language and the English language. In the first part, we describe the process of building the CzEngVallex lexicon. In the second part, based on the aligned data of the Prague Czech-English Dependency Treebank, we compare the valencies of verbal translation equivalents and comment of their differences. We classify the differences according to their underlying causes. The causes can be based in the linguistic structure of the languages, they can include translatological reasons, or they can be grounded in the character of the descriptive linguistic theory used.

<p>Previous research has shown that the type of ditransitive sentence preferred by English speakers in a particular linguistic context is significantly influenced by the relative ‘givenness’ of the post verbal arguments (i.e. the assumption that the referent of a linguistic expression is known to the speaker/hearer). This ‘givenness constraint’ has also been shown to play a role in the ditransitive sentence preferences of bilingual English speakers. Some have claimed that the realization of the givenness constraint in the ditransitive sentence preferences of bilingual English speakers is significantly influenced by the characteristics of their dominant language; however, no studies that I am aware of have explicitly compared the preferences of speakers whose dominant languages feature different sets of ditransitive sentence types, making this claim difficult to evaluate. Additionally, the effects of task type (i.e. the experimental task(s) employed by a particular study), and variables related to speakers’ experience with English and English proficiency, relative to language background are unclear. This study attempts to clarify the role of language background in the realization of the givenness constraint by recruiting three groups of English speakers: German-English bilingual speakers living in Germany, Spanish-English bilingual speakers living in Mexico, and monolingual English speakers living in the United States. The three groups completed three tasks, all of which were administered over the internet: a scalar acceptability judgement task, a forced choice task, and a self-paced reading task. The results from the two bilingual groups showed significant effects of language background, even after factors related to English proficiency and English experience were taken into account. The results support an interpretation where language background plays a significant role in the realization of the givenness constraint on bilingual speakers’ ditransitive sentence preferences. </p>

Cette étude porte sur le complément de régime (CR) des verbes espagnols, du point de vue de la variation grammaticale. Les descriptions actuelles caractérisent le CR comme une fonction syntaxique argumentative, mais aucun critère formel définitif ne permet son identification systématique. Ce complément est défini comme un groupe prépositionnel sélectionné par le prédicat. Cette définition tient compte à la fois des propriétés lexicales (la rection prépositionnelle) et syntaxiques (la forme « préposition + terme ») du CR. La similitude fonctionnelle et formelle entre le CR et d’autres fonctions syntaxiques, telles que le complément direct du verbe (CD) et le complément circonstanciel (CC), ainsi que les alternances syntaxiques et prépositionnelles impliquant le CR, reflètent la complexité de notre objet d’étude. Pour déterminer le statut grammatical et la variation (alternances prépositionnelles et de construction) du CR, nous avons conçu notre recherche à partir des données dialectales de l’espagnol contemporain. Le point de départ du projet a été la confection d’une liste de verbes prépositionnels, de laquelle nous avons tiré une sélection de huit prédicats verbaux représentatifs des différents types de verbes qui exigent un CR : confiar (‘avoir confiance en’), admirar(se) (‘admirer’), corresponder(se) (‘correspondre à’), asomar(se) (‘regarder par’), atar (‘nouer’), tachar (‘accuser de’), insistir (‘insister sur’) et consistir (‘consister à’). Avec les données obtenues nous avons aussi testé les critères syntaxiques et sémantiques d’identification du CR. Les résultats nous permettent de confirmer, d’une part, les alternances syntaxiques et prépositionnelles déjà connues, mais aussi d’autres tendances caractéristiques de l’espagnol dialectal. Nos résultats suggèrent une variation morphosyntaxique beaucoup plus complexe que celle qui avait été signalée par le passé. En général, les données montrent que le CR prend part à différents processus grammaticaux : la transitivisation (insistir en algo > insistir Ø algo ‘insister sur quelque chose > insister Ø quelque chose’), intransitivisation (admirar a alguien > admirarse de alguien ‘admirer quelqu’un > s’admirer [de] quelqu’un’) et adverbialisation (atar algo al árbol > atarlo allí ‘nouer quelque chose à l’arbre > l’y nouer’). Certains de ces processus comprennent aussi des innovations grammaticales. Bien que ces innovations semblent se limiter à certaines variétés vernaculaires (confiar en mí ~ confiar mío ‘avoir confiance en moi ~ *avoir confiance mien’, consistir {en ~ de ~ Ø} + que ‘consister {à ~ *de ~ *Ø} quelque chose’), elles reflètent quand même les possibilités du système linguistique. D’autre part, le comportement de la préposition montre des tendances liées non seulement à la variation, à la fixation ou à la désémantisation, mais aussi à la disparition en tant qu’indice fonctionnel. Immergé dans les processus de grammaticalisation susmentionnés, le CR se trouve présentement dans un statut fonctionnel « liminal », semblable non seulement à celui du CD et du CC, mais également à celui du complément indirect du verbe (confiar algo a alguien ~ confiar algo en alguien ‘confier quelque chose à quelqu’un ~ confier quelque chose en quelqu’un’) et du complément prédicatif (tacharlo de ingenuo ~ tacharlo ingenuo ‘le traiter d’idiot ~ le traiter idiot’). Nous concluons que la description actuelle du CR devrait prendre en compte les processus grammaticaux qui permettent d’expliquer la variation morphosyntaxique de cette fonction syntaxique ainsi que sa proximité fonctionnelle aux autres compléments du prédicat verbal. / This study focuses on the governed prepositional object, known in Spanish as ‘complemento de régimen verbal’ (CR), from the perspective of grammar variation. Previous descriptions have shown that, despite it being recognized as a verbal complement, no syntactic criteria have proved useful for the CR’s identification yet. When defined as a ‘prepositional phrase selected by the predicate’, the CR implies both lexical (government) and syntactic (the schema ‘preposition + phrase’) properties. Its functional and formal similarities with other complements, such as the direct object (CD) and circumstantial object (CC), as well as the syntactic alternations in which it is involved, reflect the complexity behind this phenomenon. To determine the grammatical status of this complement, as well as its variation (syntactic and prepositional alternations), I designed a corpora-based study with contemporary Spanish examples. The first step was to build a list of prepositional verbs, from which I obtained a selection of eight verbal predicates representing various types of verbs that require a CR: confiar (‘to trust’), admirar(se) (‘to admire’), corresponder(se) (‘to correspond with’), asomar(se) (‘to peek in’), atar (‘to tie’), tachar (‘to label’), insistir (‘to insist’) y consistir (‘to consist of’). The sample retrieved from the databases allowed to test the syntactic and semantic criteria used for identifying the CR. Not only do the results confirm the syntactic and prepositional alternations that were already well-known, but they also reveal some uses, in dialectal Spanish, that point to a more complex variation. In general, the data shows that the CR takes part in various grammatical processes: transitivization (insistir en algo > insistir Ø algo ‘to insist on something > to insist Ø something’), intransitivization (admirar a alguien > admirarse de alguien ‘to admire someone > to admire [oneself] of someone’), and adverbialization (atar algo al árbol > atarlo allí ‘to tie something to the tree > to tie it there’). Some of these processes are linked to grammatical innovations that, in spite of being restricted to some vernacular varieties (confiar en mí ~ confiar mío ‘to trust me ~ to trust mine’; consistir {en ~ de ~ Ø} + que ‘consist {of ~ on ~ Ø} + that’), also reflect the possibilities of the language system. Moreover, the preposition that introduces the CR is prone not only to variation, fixation, and (semantical) bleaching, but even to erasing. In these processes, the CR exhibits a ‘liminal’ functional status, similar to the CD and the CC, and sometimes even closer to the indirect object (confiar algo a alguien ~ confiar algo en alguien ‘to trust something to somebody ~ to trust something on somebody’) and the predicative complement (tacharlo de tonto ~ tacharlo tonto ‘label him as silly ~ label him silly’). In conclusion, the current grammatical accounts should consider the grammaticalization processes involving the CR that allow for an explanation of its morphosyntactic variation and functional closeness to other verbal complements. / En este trabajo se estudia el complemento de régimen verbal (CR) desde la perspectiva de la variación gramatical. Las descripciones previas han mostrado cómo, a pesar de que el CR es reconocido como una función sintáctica de naturaleza argumental, no existen criterios formales definitivos para su identificación. Su definición gramatical como grupo preposicional seleccionado por el predicado presupone la consideración de sus propiedades léxicas (la rección preposicional) a la par de sus propiedades sintácticas (la forma ‘preposición + término’). Su similitud funcional y formal con otras funciones sintácticas, como el complemento directo (CD) o el circunstancial (CC), así como las distintas alternancias en las que participa, reflejan la complejidad que hay detrás de este fenómeno. Para determinar el estatuto gramatical de este complemento y la variación que presenta (alternancias de construcción y alternancias preposicionales), se diseñó un estudio basado en datos del español actual. Se parte, para ello, de la confección de una lista de verbos preposicionales, de la cual se obtiene una selección de ocho predicados verbales representativos de los distintos tipos de verbos que requieren semánticamente un CR: confiar, admirar(se), corresponder(se), asomar(se), atar, tachar, insistir y consistir. Con la muestra obtenida, se someten a prueba, asimismo, los criterios sintáctico-semánticos de reconocimiento del CR. Los resultados del estudio permiten confirmar, por un lado, las alternancias construccionales y preposicionales ya conocidas, pero también otras tendencias propias del español dialectal que apuntan a una variación mucho más compleja. En general, los datos demuestran que el CR participa en distintos procesos gramaticales: transitivización (insistir en algo > insistir Ø algo), intransitivización (admirar a alguien > admirarse de alguien) y adverbialización (atar algo al árbol > atarlo allí). Algunos de estos procesos incluyen innovaciones gramaticales que, si bien parecen restringidas a algunas variedades vernáculas (confiar en mí ~ confiar mío; consistir {en ~ de ~ Ø} + que), reflejan las posibilidades del sistema de la lengua. Por otro lado, la preposición introductora muestra tendencias no solo a la variación, fijación y desemantización, sino incluso a la desaparición. En estos procesos de gramaticalización, el CR muestra un estatuto funcional ‘liminal’, cercano al CD y al CC, pasando por el complemento indirecto (confiar algo a alguien ~ confiar algo en alguien) y el complemento predicativo (tacharlo de ingenuo ~ tacharlo ingenuo). En síntesis, la descripción actual del CR debe tomar en cuenta los procesos gramaticales que permiten explicar su variación morfosintáctica y su cercanía funcional con otros complementos del predicado verbal.

Diese Arbeit beschäftigt sich mit den morphosyntaktischen und derivationellen Eigenschaften von Nominalisierungen im modernen Hebräisch und ihrer strukturelle Repräsentation. Eine zentrale Fragestellung im Rahmen von ‚hybriden‘ Wortbildungen wie Nominalisierungen ist die Ähnlichkeit bzw. die Unähnlichkeit zu den ihr zugrundeliegenden Verben. Unter Heranziehung des Hebräischen, einer Sprache mit reicher morphologischer Markierung, sowohl bei Verben als auch bei Nominalisierungen, werden mehrere Divergenzen zwischen Verben und entsprechenden Nominalisierungen im Bereich der Argument- und Ereignisstruktur eliminiert. Ausgehend von der einflussreichen These der Gleichsetzung von Nominalisierung und Passivierung untersucht diese Studie die syntaktische Struktur und deren Interaktion mit dem Wortbildungsprozess der Nominalisierung und zeigt, dass Eigenschaften, die für Passivformen typisch sind, in Nominalisierungen fehlen. Dabei präsentiert diese Studie mit der Untersuchung morphosyntaktischer Faktoren und deren Beziehungen zu Nominalisierungen, der Inkonsistenzen aufzeigt. Durch einen Vergleich von etwa 3000 Verben auf Basis der Verbklassenmorphologie ergibt sich eine signifikante Asymmetrie zwischen Nominalisierungen, die eine mediale/intransitive Markierung tragen, und Nominalisierungen, die als aktiv markiert sind, wobei sich die mediale Form in zwei klar definierten syntaktischen Kontexten als weniger produktiv erweist. Dies zeigt sich auch dadurch, dass alternierende Wurzeln, also Wurzeln die sowohl aktive als auch mediale Verbformen ausbilden können, bilden ihre Nominalisierungen auf Basis ihrer aktiven Form. Auf Basis der Konzepte von Konkurrenz und Markiertheit werden diese paradigmatischen Lücken nicht als grammatisch bedingte Inkompatibilitäten analysiert, sondern als eine generelle Präferenz für weniger markierte Formen (aktiv-markierte Nominalisierungen) gegenüber komplexeren (medial-markierte Nominalisierungen), wie in der Performanz häufig zu beobachten. / This study is concerned with the properties, structural representation and derivational patterns of deverbal nouns (DNs) in Modern Hebrew. A recurring question arises in the context of such ‘hybrid’ formations: precisely how similar or far-apart are these derivatives from the verbs from which they originate? Enlisting Hebrew, a language with rich morphological marking on both verbs as well as DNs, several loci of divergence between verbs and respective DNs in the domain of argument- and event-structure are eliminated. Taking as a point of reference the influential view which equates the processes of nominalization and passivization, this study scrutinizes syntactic structure and its interaction with nominalization, showing that behaviours typical of passives are absent from DNs. a finding which weakens long-standing beliefs bearing on this class. A novel area of exploration offered in this study is the examination of morpho-syntactic factors and their interaction with nominalization, a domain where inconsistencies do arise. What emerges from a comparison of some 3000 verbs based on verb-class (templatic) morphology is a significant asymmetry between DNs carrying Middle (intransitive) marking and DNs marked as Active, wherein Middle forms are found to be less productive in two well-defined syntactic contexts. Not entirely absent, however, the same roots which fail to surface with Middle morphology are perfectly licit when derived from the corresponding Active verb (in case of alternating roots). Building on the notions of competition and markedness, such paradigmatic gaps are analysed not as grammatically-determined incompatibilities, but as a consistent preference for less-marked forms (Active-marked DNs) over more complex ones (Middle-marked DNs), a trend which lies within the realm of performance. As such, Hebrew DNs constitute a case study of the interrelations between the syntactic and morphological modules, and pragmatics.