Generative Lexicon

Resumen En el siguiente trabajo se estudian tres expresiones idiomáticas de tipo locución verbal (LV) con un componente o lexía relacionada con la gastronomía del español de Venezuela. Las expresiones se extrajeron de un corpus compuesto por 207 expresiones idiomáticas obtenidas mediante entrevista informales realizadas a hablantes caraqueños. Las expresiones se analizaron a partir de la Teoría del Lexicón Generativo y la teoría de la metáfora. El análisis de las expresiones da cuenta de que en ellas opera un proceso metafórico en el cual los alimentos toman el lugar que usualmente ocuparía una persona en el dominio de origen.

Generative Lexicon is a theory of linguistic semantics which focuses on the distributed nature of compositionality in natural language. Unlike purely verb-based approaches to compositionality, Generative Lexicon (henceforth, GL) attempts to spread the semantic load across all constituents of the utterance. Central to the philosophical perspective of GL are two major lines of inquiry: (1) How is it that we are able to deploy a finite number of words in our language in an unbounded number of contexts? (2) Is lexical information and the representations used in composing meanings separable from our commonsense knowledge?

En esta versión publicada de la tesis se estudian los fenómenos morfológicos (formación aspectual en ruso) y sintácticos (estructuras medias, pasivas e inacusativas, muy especialmente dentro del grupo de los verbos de movimiento) en los que influyen las propiedades léxicas de las palabras. El trabajo se inscribe dentro del modelo del Lexicón Generativo. Palabras clave: aspecto, Aktionsart, estructura eventiva, estructura de qualia, mecanismos composicionales, ligamiento selectivo, co-composición, coacción.

This paper offers a scalar analysis of Russian verbal forms derived with the prefix pro-. The proposed approach is crucially based on two notions integrated into the Generative Lexicon framework: 'Dynamic Event Structure', a representation of the internal makeup of events in terms of subevents or phases which allows tracking the change of the arguments’ properties over time; and 'change function', which accounts for the different ways in which these properties may be affected (i.e., initiated, terminated, modified or left unchanged) in the course of the event. Special emphasis is placed on the role of scalar properties of the nominal arguments in the composition of change-of-state predicates.

Relevance of Qualia Relations in Coercive Contexts Elisabetta Ježek Università di Pavia Strada Nuova 65 27100 Pavia, Italy jezek@ unipv. it Valeria Quochi Istituto di Linguistica Com-putazionale, CNR Via Moruzzi, 1 56100, Pisa, Italy Valeria. quochi @ ilc. cnr. it Nicoletta ...

In this paper, I discuss the notion of normality in language from the viewpoint of default interpretations: that is, salient interpretations which arise under normal circumstances and in the absence of evidence to the contrary. While the concept of default interpretations is largely unproblematic from an intuitive point of view, it is by no means well-defined, which has resulted in a plethora of accounts attempting to explain the source of these 'fast and frugal' interpretations. Using English pre-nominal possessive phrases (definite expressions of the kind John's book, my mother, yesterday's newspaper, etc.) as an example , I examine a type-based account attributing this source to the lexical semantics of their parts and discuss in particular the question of whether default interpretations are dependent on context for their derivation. I present an empirical investigation into possessive interpretations which uses the absence of linguistic context as a proxy for inter-pretability and find that type-based approaches lack in explanatory power when it comes to accounting for the broad range of interpretations elicited. In their stead, I propose a more pragmatic notion of default interpretations which affords them the influence of addressee-dependent salience considerations culled from the extra-linguistic context and encyclopaedic assumptions activated when certain kinds of linguistic material are encountered .

In this paper we examine the role that compositional mechanisms and lexical semantics play in the determination of informativeness at the phrasal and clausal level. While the computation of the “relevance” of an utterance is largely determined by pragmatic factors (such as quantity), we argue that phrasal informativeness can, in many cases, be computed
compositionally and independently of pragmatics. To illustrate this, we focus on the well-documented contrast between
predicative and derived participial modificational constructions in English ('build a house' results in well-formed sentences,
while *'a built house' does not). In our analysis, informativeness within an NP is computed in terms of minimal model generation (Blackburn and Bos, 2008), using the semantics associated with the qualia of the head noun; that is, modification is informative whenever a qualia value is not satisfied in all models.

Abstract. The long-term historical development and broad geographical variation of Chinese character (Hanzi/Kanji) has made it a crosscultural information sharing platform in East Asia. However, due to the lack of proper research framework, the integration of heterogeneous knowledge grounded in Hanzi and its variants has been a thorny problem. In this paper, we propose a theoretical framework for the knowledge representation of Hanzi in the cross-cultural context. Our proposal is mainly based on two resources: Hantology and ...

The availability of computerized lexicons, thesauri and “ontologies”—we discuss this terminology—makes it possible to formalize semantic aspects of information as used in the analysis, design and implementation of information systems (and in fact general software systems) in new and useful ways. We survey a selection of relevant ongoing work, discuss different issues of semantics that arise, and characterize the resulting computerized information systems, called CLASS for Computer-Lexicon Assisted Software Systems. The need for a “global” common ontology (lexicon, thesaurus) is conjectured, and some desirable properties are proposed. We give a few examples of such CLASS-s and indicate avenues of current and future research in this area. In particular, certain problems can be identified with well-known existing lexicons such as CYC and WordNet, as well as with sophisticated representation-and inference engines such as KIF or SHOE. We argue nevertheless that large public lexicons should be simple, i.c. their semantics become implicit by agreement among “all” users, and ideally completely application independent. In short, the lexicon or thesaurus then becomes the semantic domain for all applications.

Thomas & Ernout (1964), Asencio (1998) e Baños Baños (2009) descrevem e analisam os valores de cum+ablativo, salientando um valor temporal em dados de Plauto como: "Egone aps te abii hinc hodie cum diluculo?". Porém, este valor da preposição é o menos revisado na literatura, sendo que ´cum´ veicula mais frequentemente companhia, modo e circunstância. Levando em consideração esse cenário, este trabalho aborda o sintagma preposicional latino introduzido pela preposição ´cum´, no contexto em que se apresenta com um valor temporal. Assim, objetivamos lançar luz sobre o particular contexto de aparição do sintagma preposicional cum+ablativo. Metodologicamente, baseamo-nos no enfoque lexical-semântico do modelo do Léxico Gerativo de Pustejovsky (1995) que contribui, por sua vez, à explicação da gama de valores que pode adquirir a preposição em contextos discursivos determinados.

Este trabajo busca contribuir a la descripción del sintagma preposicional latino introducido por cum en contextos en que presenta un valor temporal. Exploramos la posibilidad de aplicar el modelo de la Teoría del Lexicón Generativo de Pustejovsky (1995) para explicar los rasgos subléxicos y los mecanismos generativos que habilitan este valor en el SP. Nuestra propuesta de análisis sostiene que el valor temporal es producto de la acción conjunta, generada por el mecanismo de co-composición, de los rasgos [–dinamismo], [–orientación] y [+coexistencia] de la estructura de qualia que presentamos para la preposición cum y los rasgos subléxicos de los nombres eventivos que la preposición toma como complemento.

This paper examines the construction of verbal aspect as a sub-case of generation of verbal lexical meaning. The analysis is based on Pustejovsky's (1995) Generative Lexicon and assumes the existence of a set of lexical agreement processes that match the lexical features of arguments (and adjuncts) with the information contained in the meta-entry of the verb in the mental lexicon. Verbal meta-entries include distinct sub-structures, among them the Event Structure, which is composed of different phases or sub-events. The materializa-tion of some subevents or others depends on the agreement processes triggered by the arguments and adjuncts and gives rise to different aspectual meanings for the same verb. 1. The construction of verbal aspect. General remarks The construction of verbal aspect is a complex phenomenon wherein the different types of information provided by words making up the predicate are involved. This interaction, which was considered as problematic in classical literature on aspect, was attributed to the compositional nature of lexical aspect. Recently, it has been described as the result of the functioning of mechanisms responsible for generating the lexical meaning. Specifically, it is assumed that there is a mechanism of aspectual coercion whereby argument information determines the lexical aspect of verbs.

The most productive way to encode ablative, privative, and reversative
meanings in current Catalan and Spanish is by means of des- prefixation. This paper investigates how these related values are obtained both from a structural and from a conceptual perspective. To analyze the structural behaviour of these predicates, a new neoconstructionist model is adopted: Nanosyntax, according to which lexical items are syntactic constructs. As for the conceptual content associated to these verbs, it is accounted for by means of a non-canonical approach to the Generative Lexicon Theory developed by Pustejovsky (1995 ff.). The core proposal is that des- prefixed verbs with an ablative, a privative, or a reversative value share the same syntactic structure, and that the different interpretations emerge as a consequence of the interactions generated, at a conceptual level, between the Qualia Structure of the verbal root and that of the internal argument of the verb.

Over the last fifteen years, following Perlmutter and Postal's (1983) work, there has been constant interest in the problem raised by transitivity alternations. Perlmutter and Postal argued that the single argument of intransitive verbs may either be generated as an object ...

ABSTRACT: The relation between the generative and interpretative views of lexical meaning hasn't been fully explored in current theoretical models about polysemy. The Generative Lexicon (Pustejovsky 1995, 2006, 2011) proposes that lexical-grammatical items operate mechanisms on the word semantic structure in order to generate one particular sense in context. Here, we use the term sense generation indicator (SGI) to refer to these elements that operate the generative mechanisms and, at the same time, are textual evidence that allow to interpret the word in context. This work aims to identify and analyze SGI types and to describe them. We study them in a textual corpus of nouns that presents regular polysemy between a locative sense and a non-locative sense. The following hypotheses are tested: i) the items that operate the generative mechanisms are of two kinds: specific for certain senses, and general to the whole nominal category, each having a different " strength " for disambiguation; (ii) the interaction between SGI with different strengths is a decisive factor for the perception of the nouns as ambiguous and this supports the dynamic quality of the lexical meaning; (iii) regarding the general SGI, from the viewpoint of the generation of lexical meaning, the expression in context of a certain lexical relation is relevant, while from the methodological perspective of disambiguation it is the particular lexical item used in a particular linguistic context the one that enables a particular reading (i.e., certain hypernym, co-hyponym, or meronym).

In this paper we describe the conception of a software toolkit designed for the construction, maintenance and collaborative use of a Generative Lexicon. In order to ease its portability and spreading use, this tool was built with free and open source products. We eventually tested the toolkit and showed it filters the adequate form of anaphoric reference to the modifier in endocentric compounds.

One central issue in the generative lexicon approach is how to account for the context dependence of certain aspects of the meanings of lexical items. Pustejovsky (1995, 2005), Ginzburg (to appear) and Ginzburg and Cooper (2004) exemplify how the interpretation of ...

T-Lex is a thesaurus manually constructed from a traditional dictionary, adopting aspects from the Generative Lexicon Theory (GLT) proposed by James Pustejovsky, as well as object-orient ed software modeling concepts. The Qualia structure of GLT is adapted for describing nouns, verbs and adjectives on a differentiated way, respectively, as objects, operations, and states. Generative operations which return sets of related

In this paper we describe a range of techniques used at NMSU CRL for accelerating the development of MT systems. These techniques enable semi-automatic development of a number of components of a multilingual MT system, thereby enabling rapid deployment of MT capabilities in a new language. First, we describe the core multi-engine, multilingual architecture that enables the different techniques to