Thèses sur le sujet « Linguistic formal model »

2010 - 2011
The principal aim of this research is describing to which extent formal models for linguistic data structuring are crucial in Natural Language Processing (NLP) applications. In this sense, we will pay particular attention to those Knowledge Management Systems (KMS) which are designed for the Internet, and also to the enhanced solutions they may require. In order to appropriately deal with this topics, we will describe how to achieve computational linguistics applications helpful to humans in establishing and maintaining an advantageous relationship with technologies, especially with those technologies which are based on or produce man-machine interactions in natural language. We will explore the positive relationship which may exist between well-structured Linguistic Resources (LR) and KMS, in order to state that if the information architecture of a KMS is based on the formalization of linguistic data, then the system works better and is more consistent. As for the topics we want to deal with, frist of all it is indispensable to state that in order to structure efficient and effective Information Retrieval (IR) tools, understanding and formalizing natural language combinatory mechanisms seems to be the first operation to achieve, also because any piece of information produced by humans on the Internet is necessarily a linguistic act. Therefore, in this research work we will also discuss the NLP structuring of a linguistic formalization Hybrid Model, which we hope will prove to be a useful tool to support, improve and refine KMSs. More specifically, in section 1 we will describe how to structure language resources implementable inside KMSs, to what extent they can improve the performance of these systems and how the problem of linguistic data structuring is dealt with by natural language formalization methods. In section 2 we will proceed with a brief review of computational linguistics, paying particular attention to specific software packages such Intex, Unitex, NooJ, and Cataloga, which are developed according to Lexicon-Grammar (LG) method, a linguistic theory established during the 60’s by Maurice Gross. In section 3 we will describe some specific works useful to monitor the state of the art in Linguistic Data Structuring Models, Enhanced Solutions for KMSs, and NLP Applications for KMSs. In section 4 we will cope with problems related to natural language formalization methods, describing mainly Transformational-Generative Grammar (TGG) and LG, plus other methods based on statistical approaches and ontologies. In section 5 we will propose a Hybrid Model usable in NLP applications in order to create effective enhanced solutions for KMSs. Specific features and elements of our hybrid model will be shown through some results on experimental research work. The case study we will present is a very complex NLP problem yet little explored in recent years, i.e. Multi Word Units (MWUs) treatment. In section 6 we will close our research evaluating its results and presenting possible future work perspectives. [edited by author]
X n.s.

Orientador: Linda Gentry El-Dash
Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem
Made available in DSpace on 2018-08-06T06:03:50Z (GMT). No. of bitstreams: 1 Sousa_DenilsonAmade_M.pdf: 683113 bytes, checksum: b56e92e2cf7ccf5d2606b65b00c15c99 (MD5) Previous issue date: 2006
Resumo: Este estudo realizou uma descrição de um fenômeno lingüístico chamado alternância das formas dativas em inglês (dative alternation) e verificou, através de um teste de julgamento de gramaticalidade, se professores e alunos brasileiros de inglês têm conhecimento implícito ou explícito acerca deste fenômeno. Comparando a intuição de nativos americanos com a intuição de alunos e professores brasileiros, chegou-se à conclusão de que nem os professores nem os alunos possuem esse conhecimento. A partir desses resultados foram realizadas intervenções em sala de aula com o intuito de fazer com que aprendizes de inglês de nível intermediário adquirissem a alternância das formas dativas. Essas intervenções foram baseadas em processamento de insumo e instrução para processamento (como formulados por VanPatten, 1993,1996) e tiveram um design de pré-teste/pós-teste. As análises estatísticas desse segundo experimento ¿ intervenção ¿ mostraram, com uma diferença significativa ao nível de 5%, que o grupo que sofreu o tratamento baseado em instrução para processamento de insumo aprendeu a alternância das formas dativas apenas no nível imediato, já que essa aprendizagem se perdeu depois de passado um mês. As análises estatísticas referentes ao grupo que sofreu um tratamento baseado numa abordagem tradicional indicaram que não houve aprendizagem imediata estatisticamente relevante. Essas análises também apontaram, com uma diferença significativa ao nível de 10%, para a perda do conhecimento, estatisticamente não significativo, adquirido na intervenção. O conhecimento do grupo de controle ¿ aquele que não sofreu nenhum tipo de tratamento ¿ se manteve o mesmo durante o período da pesquisa. Evidenciou-se, assim, que, pelo menos para esse tipo de estrutura, a Instrução para Processamento de Insumo surtiu um efeito apenas imediato, que não durou muito tempo. Essa descoberta é interessante, pois se pode imaginar o motivo pelo qual isso tenha ocorrido e conseqüentemente se especular acerca daquilo que pode ser feito para se evitar a perda do conhecimento adquirido na intervenção
Abstract: This study was designed to describe a linguistic phenomenon in English known as dative alternation. A grammaticality judgment task was used to ascertain whether Brazilian teachers and learners of English have any knowledge of that phenomenon, either implicit or explicit. When the intuition of native Americans was compared with that of Brazilian teachers and learners, it was attested that neither the teachers nor the students possess that kind of knowledge. Based on these results, two kinds of classroom intervention were carried out in an attempt to investigate the acquisition of the possession constraint underlying dative alternation by intermediate level English students. These interventions, based on input processing and processing instruction (as formulated by VanPatten, 1993, 1996), involved a pre-test/post-test design. The statistical analysis of this experiment showed, with a level of significance of 5%, that the group submitted to processing instruction learned the restriction underlying dative alternation, but only in the short term, since this learning was lost after a month. The statistical analysis of the group submitted to a treatment involving a traditional approach did not reveal any statistically significant kind of short term learning. Although the analysis suggested, with a level of significance of 10%, that any knowledge acquired during intervention was lost. The knowledge of dative alternation of the control group ¿ a group which did not undergo any kind of treatment ¿ remained unaltered. Thus, it was demonstrated that, at least for this kind of structure, processing instruction had only a short term effect, which did not last as long as a month. These findings are interesting because one may wonder why this happened and, consequently, speculate about what can be done in order to avoid the loss of the knowledge acquired through the processing instruction intervention
Mestrado
Mestre em Linguística Aplicada

Orientadores: Ruth Elisabeth Vasconcellos Lopes, Charlotte Marie Chamberlland Galves
Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem
Made available in DSpace on 2018-08-24T00:21:30Z (GMT). No. of bitstreams: 1 Faria_Pablo_D.pdf: 2488292 bytes, checksum: c75ee78601e9b73c790d12ce7e2d414e (MD5) Previous issue date: 2013
Resumo: Neste trabalho, o fenômeno de aquisição de uma língua natural é investigado através de uma modelagem computacional. O aprendiz modelado - apelidado de IASMIM - se caracteriza como um modelo computacional integrado de aquisição de primeira língua, visto que integra os processos de aquisição lexical e sintática. Além disso, o modelo foi concebido de modo a atender certos critérios de plausibilidade empírica e psicológica. A perspectiva teórica que norteia a investigação é a da Gramática Gerativa (cf. Chomsky, 1986) e este é um modelo voltado para a competência linguística, não um modelo de processamento ou de performance (i.e., de uso do conhecimento linguístico). O aprendiz modelado é capaz de adquirir um conhecimento gramatical relativamente abrangente e demonstra algum potencial translinguístico, particularmente no que diz respeito a variações de ordem. As simulações para avaliação do modelo permitem observar a emergência de padrões de adjunção e de recursividade na gramática, considerados aqui como as principais evidências de um conhecimento sintático mais elaborado. Finalmente, o modelo incorpora algumas noções caras à teoria sintática no âmbito do Programa Minimalista (cf. Chomsky, 1995b), tais como set- Merge, pair-Merge, "traço seletor" (cf. Chomsky, 1998), em conjunto com assunções sobre a binariedade das representações sintáticas e a hipótese de que a ordem linear não tem papel na sintaxe (cf. Uriagereka, 1999). O modelo incorpora, ainda, uma versão da representação semântico-conceitual proposta em Jackendoff (1990). Nesta modelagem, estas noções e assunções ganham uma interpretação concreta e integrada, interagindo na determinação das propriedades do conhecimento adquirido
Abstract: In the present work, the acquisition of natural languages is investigated through a computer simulation. The modelled learner - dubbed IASMIM - is characterized as an integrated computational model of first language acquisition, in the sense that it integrates the processes of lexical and syntactic acquisition. Furthermore, the model was conceived in order to be empirically and psychologically plausible. The theoretical perspective of this enterprise is that of Generative Grammar (cf. Chomsky, 1986) and this is a model concerned with linguistic competence, rather than language processing or performance (i.e., how the acquired knowledge is put to use). The modelled learner is capable of acquiring a relatively broad grammatical knowledge and shows some crosslinguistic abilities, in particular, the ability to handle languages with distinct word orders. In the simulations for evaluation of the model we can observe the emergence of adjunction and recursive patterns in the grammar, taken here as the main pieces of evidence of a more elaborated syntactic knowledge. Finally, the model embodies some central notions for syntactic theory under the Minimalist Program (cf. Chomsky, 1995b), such as set-Merge, pair-Merge and "selector feature" (cf. Chomsky, 1998), together with the assumptions that syntactic representations are strictly binary branching and that linear word order has no significant role in syntactic phenomena (cf. Uriagereka, 1999). The model also embodies a version of the semantic-conceptual representation proposed in Jackendoff (1990). They take a concrete and integrated existence in this model, interacting with one another to determine the properties of the acquired grammatical knowledge
Doutorado
Linguistica
Doutor em Linguística

2013 - 2014
The main aim of this research is to propose a structuralist approach for knowledge processing by means of ontology learning and population, achieved starting from unstructured and structured texts. The method suggested includes distributional semantic approaches and NL formalization theories, in order to develop a framework, which relies upon deep linguistic analysis... [edited by author]
XIII n.s.

Orientador: Claudia Thereza Guimarães de Lemos
Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem
Made available in DSpace on 2018-08-08T08:59:15Z (GMT). No. of bitstreams: 1 Lodovici_FlaminiaManzanoMoreira_D.pdf: 632382 bytes, checksum: 219bf2f628c8e30507068fba919cc62f (MD5) Previous issue date: 2007
Resumo: Este trabalho de investigação objetiva abordar a natureza da questão idiomática e a relação que, por meio dela, reclama reflexão: o jogo entre o universal do funcionamento lingüístico e as manifestações singulares de fala. A análise deixa ver que o idiomatismo é mesmo um fato de língua; resulta de um jogo combinatório no qual seu efeito de unidade se configura, e delimita-se no fluxo da fala "em ato". Não se pode propriamente prever seu sentido antes que ele se atualize numa determinada fala. A hipótese é a de que "expressões idiomáticas" se comportam como outras quaisquer no sistema: elas também estão sujeitas a operações de composição/ recomposição que renovam seu modo de fazer presença nos dizeres. Mesmo em se levando em consideração que sentidos relativos a enunciados idiomáticos podem se estabilizar no uso, o que importa é que estabilização não é sinônimo de fixidez. A investigação indica que essas "formas de significar" comportariam duplo sentido. A princípio, a tendência deste trabalho foi tomar essa característica como diferencial dos outros elementos da língua. Mas o que se impõe é reconhecer que esse é mesmo o destino de todo o dizer. Por fim, o idiomatismo exibe uma diferença, uma extravagância que parece vincular-se à sua natureza essencialmente metafórica que, por sua vez, responde pela produção de um, muitas vezes, inesperado efeito significativo
Abstract: The purpose of this study is to approach the nature of the idiomatic issue and the relation that requires reflexion: the play between the universal aspect of linguistic functioning and the unique manifestations of speech. This analysis allows you to see that idioms are actually a speech fact; they result from a combinatory play in which their unity effect is configured and bound in the speech flow in actu [en acte/in act]. Their meaning cannot be foreseen before they are materialized in a specific speech. The hypothesis is that ¿idiomatic expressions¿ behave like any others in the system: they are also subject to composition / recomposition operations which renew the manner in which they are present in the utterances. Even if we consider that senses concerning idiomatic utterances can be stabilized in usage, what really matters is that stabilization is not synonymous with immutability. Research shows that such forms of meaning attribution allow for double meaning. At first, this study was inclined to consider this as a distinguishing feature with respect to the other features of speech. But what must be acknowledged is that all utterances have the same fate. Finally, idioms show a difference, an extravagancy that seems to be linked to their essentially metaphorical nature which, in turn, is responsible for the production of a, very often, unexpected signifying effect
Doutorado
Linguistica
Doutor em Linguística

Dietary health risks are among the lifestyle-related health challenges seen all over the world, and are connected to every individual's daily behavior. They have motivated a shift from reactive healthcare to proactive health communication and promotion, and therefore have attracted the increased interest of communication professionals and researchers. Salt consumption has become a major dietary risk in the current world, which has been linked to a variety of noncommunicable diseases. China has the world’s third highest mortality rate caused by a diet high in sodium while Australia has made remarkable strides in controlling salt consumption. Exposure to persuasive health communication has been viewed as a significant strategy to influence people's beliefs, attitudes, intentions, or behaviors with the goal of public health intervention. In this regard, the overall purpose of this research is to investigate and compare the linguistic characteristics of health news between the comparable and bilingual Australian-Chinese mass media news corpora. The comparable corpora are self-built, compiled with health-themed news reporting on salt consumption. By analyzing the corpora, this study analyzes communication characteristics from two perspectives, namely information evidentiality and relevance. Analysis will reveal how health messages are presented to be persuasive and effective, which are integral to public health issues and risk perception. Through the application of corpus linguistics and computational linguistics techniques, the similarities and differences of linguistic characteristics will be revealed. This study attempts to advance comparable health communication research. The research findings will send a critical message to communication professionals that health news has the potential to change people’s risk perceptions. This will hopefully have broader implications for the improvement of health news quality.

The aim of distributional semantics is to design computational techniques that can automatically learn the meanings of words from a body of text. The twin challenges are: how do we represent meaning, and how do we learn these representations? The current state of the art is to represent meanings as vectors - but vectors do not correspond to any traditional notion of meaning. In particular, there is no way to talk about 'truth', a crucial concept in logic and formal semantics. In this thesis, I develop a framework for distributional semantics which answers this challenge. The meaning of a word is not represented as a vector, but as a 'function', mapping entities (objects in the world) to probabilities of truth (the probability that the word is true of the entity). Such a function can be interpreted both in the machine learning sense of a classifier, and in the formal semantic sense of a truth-conditional function. This simultaneously allows both the use of machine learning techniques to exploit large datasets, and also the use of formal semantic techniques to manipulate the learnt representations. I define a probabilistic graphical model, which incorporates a probabilistic generalisation of model theory (allowing a strong connection with formal semantics), and which generates semantic dependency graphs (allowing it to be trained on a corpus). This graphical model provides a natural way to model logical inference, semantic composition, and context-dependent meanings, where Bayesian inference plays a crucial role. I demonstrate the feasibility of this approach by training a model on WikiWoods, a parsed version of the English Wikipedia, and evaluating it on three tasks. The results indicate that the model can learn information not captured by vector space models.

Made available in DSpace on 2015-05-14T12:42:44Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 11232415 bytes, checksum: 1c66cf270d76532c9a556cb9ced8ce30 (MD5) Previous issue date: 2012-03-30
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
The main objective of this research is to develop a theoretical discussion about the phenomenon of social (re)categorization and (de)legitimation of discourse, in which we take into account social, historical, cultural and cognitive aspects involved in the use of language in times of social conflict. Bearing this objective in mind, we base this study on, at least, three different fields of Linguistics, namely, Critical Discourse Analysis and its studies on (de)legitimation of discourse, Text Linguistics and its contribution with its studies on referentiation and Cognitive Linguistics and the theory of Idealized Cognitive Models. This integration of theories aims at investigating the way the personal pronoun forms of the subjective case turn it feasible to create and reflect the several social categories which may derive from the same pronoun. Besides, we also aim at investigating the role of this linguistic resource in the phenomenon of (de)legitimation of discourse. The socio-historical-cultural moment chosen to study these issues was the Civil Rights of the sixties in the United States of America. The corpus is composed of the speeches I have a dream by Martin Luther King, Jr. and The ballot or the bullet by Malcolm X. In our analysis, we investigate how the two leaders constructed their speeches around (de)legitimation of discourse and of social (re)categorization before the racial conflict the American society faced. We highlight the relation form-meaning established by the pronominal forms as cognitively motivated, lying on beliefs, values and world knowledge shared with and learned from other individuals within the socio-cultural context in which they are inserted.
Nosso objetivo principal ao desenvolvermos esta pesquisa é estabelecermos uma discussão teórica em torno da (re)categorização social e (des)legitimação de discursos, considerando aspectos sociais, históricos, culturais e cognitivos envolvidos no uso da linguagem em momentos de conflito social. Com este objetivo em mente, baseamos este estudo em, pelo menos, três ramos da linguística, a saber, a Análise Crítica do Discurso e seus estudos sobre (des)legitimação de discursos, a Linguística Textual e sua contribuição sobre o processo de referenciação e a Linguística Cognitiva com sua teoria dos Modelos Cognitivos Idealizados. Esta integração de teorias visa a investigar o modo como a utilização de formas pronominais pessoais do caso sujeito viabiliza a criação de e reflete as várias categorias sociais que emanam de um mesmo pronome. Além disso, investigamos o papel deste recurso linguístico na (des)legitimação de discursos. O momento sóciohistórico- cultural que escolhemos para refletirmos sobre estas questões foi o Movimento por Direitos Civis dos anos sessenta nos Estados Unidos da América e o corpus selecionado é composto pelos discursos I have a dream de Martin Luther King, Jr. e The ballot or the bullet de Malcolm X. Em nossa análise, investigamos como os dois líderes negros construíram seus discursos em torno da (des)legitimação de discursos e da (re)categorização social diante do conflito racial vivido na época. Retomamos a questão de que os pronomes estabelecem uma relação forma-significado, cognitivamente motivada, baseada em crenças, valores e conhecimentos de mundo partilhados e aprendidos com outros no contexto sóciocultural em que se encontram inseridos.

This study examines the influence of formal L2 English instruction (FEI) onpronunciation accuracy of similar and new vowels in the framework of Flege’s SpeechLearning Model (Flege, 1995a, 1999, 2002), focusing on selected segmentalpronunciation features produced by native German speakers. A spectral vowelproduction experiment was carried out with 24 German students who attend sixth, ninthand twelfth grade of a south German secondary school by the time this study wasconducted. Spectral characteristics of the similar vowel /ɛ/ and the new vowel /æ/,produced by each subject of the three groups (G6, G9, and G12), were compared tovowel data of native English speakers. In order to test the influence of formalinstruction on selected phonetic segments, which are known to cause difficulties fornative German learners of English, three native English listeners judged on a three pointscale the intelligibility of the segments produced. The results of the vowel experimentshowed significant influence of FEI on pronunciation accuracy of some characteristicsof the similar vowel /ɛ/ but not of the new vowel /æ/. The listener rating experimentshowed that FEI had on average no influence on pronunciation accuracy of the three L2English learner groups.

Automatic translation has seen tremendous progress in recent years, mainly thanks to statistical methods applied to large parallel corpora. Transductions represent a principled approach to modeling translation, but existing transduction classes are either not expressive enough to capture structural regularities between natural languages or too complex to support efficient statistical induction on a large scale. A common approach is to severely prune search over a relatively unrestricted space of transduction grammars. These restrictions are often applied at different stages in a pipeline, with the obvious drawback of committing to irrevocable decisions that should not have been made. In this thesis we will instead restrict the space of transduction grammars to a space that is less expressive, but can be efficiently searched. First, the class of linear transductions is defined and characterized. They are generated by linear transduction grammars, which represent the natural bilingual case of linear grammars, as well as the natural linear case of inversion transduction grammars (and higher order syntax-directed transduction grammars). They are recognized by zipper finite-state transducers, which are equivalent to finite-state automata with four tapes. By allowing this extra dimensionality, linear transductions can represent alignments that finite-state transductions cannot, and by keeping the mechanism free of auxiliary storage, they become much more efficient than inversion transductions. Secondly, we present an algorithm for parsing with linear transduction grammars that allows pruning. The pruning scheme imposes no restrictions a priori, but guides the search to potentially interesting parts of the search space in an informed and dynamic way. Being able to parse efficiently allows learning of stochastic linear transduction grammars through expectation maximization. All the above work would be for naught if linear transductions were too poor a reflection of the actual transduction between natural languages. We test this empirically by building systems based on the alignments imposed by the learned grammars. The conclusion is that stochastic linear inversion transduction grammars learned from observed data stand up well to the state of the art.

In Chilean Spanish, second-person address is non-uniform in that the vos competes with the conventional tuteo and a third, mixed form has emerged. To add to this complexity, the form speakers choose has been shown to correspond to socioeconomic strata. Upper classes use tú, lower classes use vos, and young, middle class speakers choose the mixed form in which the verb is conjugated according to the voseo and is used with the pronoun tú. The causes and effects of this second-person schism in Chile are explored here, as well as the resulting sociolinguistic issues and consequences. In a study of printed media, television and interviewed informants, an attempt is made to confirm and validate the complexity of address in Chilean Spanish and determine the degree of the mixed voseo‟s pervasion into the mainstream.

La present tesi consisteix en una anàlisi compativa de tres gèneres de discurs ?conversa informal (llengua oral prototípica), classe magistral (gènere intermedi) i prosa acadèmica (llengua escrita prototípica)? en català, en tres aspectes de l'estructuració i la complexitat lingüística: la densitat lèxica, la composició oracional i la connexió textual (total: 90 trets lèxics, gramaticals i textuals analitzats). Els objectius principals són a/ dotar la lingüística catalana d'una descripció detallada, fins ara inexistent, del contrast lingüístic entre oralitat i escriptura, i b/ fer una aportació a la recerca internacional sobre el tema per intervenir, d'una banda, en el debat originat sobre la complexitat sintàctica i, especialment, sobre la presència d'oracions subordinades en els dos modes de discurs i, de l'altra, per ampliar aquesta mena d'investigació a l'àmbit del discurs, com han reclamat alguns tractadistes.
La presente tesis consiste en un análisis comparativo de tres géneros de discurso ?conversación informal (lengua oral prototípica), clase magistral (género intermedio) y prosa académica (lengua escrita prototípica)? en catalán, en tres aspectos de la estructuración y la complejidad lingüística: la densidad léxica, la composición oracional y la conexión textual (total: 90 rasgos léxicos, gramaticales y textuales analizados). Los objetivos principales son a/ dotar la lingüística catalana de una descripción detallada, hasta ahora inexistente, del contraste lingüístico entre oralidad y escritura, y b/ hacer una aportación a la investigación internacional sobre el tema para intervenir, por un lado, en el debate originado sobre la complejidad sintáctica y, especialmente, sobre la presencia de oraciones subordinadas en los dos modos de discurso y, por otro lado, para ampliar este tipo de investigación al ámbito del discurso, como han reclamado algunos tratadistas.
The present work consists of a comparative analysis of three discourse genres --informal conversation (prototypical oral language), academic lecture (intermediate genre), and academic prose (prototypical written language)-- in Catalan, focusing on three aspects of linguistic structure and complexity: lexical density, sentence composition, and textual connexion (total: 90 lexical, grammatical, and textual features analysed). The main purposes are: (a) to provide Catalan linguistics with a detailed description --non-available so far-- of the linguistic opposition between literacy and speech; and (b) to make a contribution to the international research in this field in order to, on the one hand, participate in the discussion about the syntactic complexity, and, especially, about the presence of subordinated clauses in the two modes of discourse, and, on the other, extend this kind of research to the discourse level, as some researchers have claimed.

We present a declarative, generative model of simple narratives. Our model contributes to automation of natural language processing because it provides a model of the relationship between surface text and its meaning. This work surpasses previous story modeling work in three important ways. First, we build on developments in commonsense knowledge representation to construct precise formal definitions of components of simple narratives and relations among these components. Second, we develop a new theory of rational intention in autonomous agents to formalize the relation of a character's mental state to actions it undertakes in a narrative. Third, we implement this model in the first story generation system to use a formal grammar for stories The centerpiece of our model is a story grammar expressing characteristics of event sequences which, when reported in natural language, constitute a narrative. This grammar draws on a separate theory of rational intention in autonomous agents. The grammar makes use of interchangeable world models specifying characters that may appear in a story, emotions they feel, actions they take, and events that happen in narratives We implement our model in a story generation system which simultaneously generates text and meaning representations. This system represents a significant achievement since it is the first such system constructed from an explicit, formal model for stories. The implementation serves as a concrete demonstration of the computational viability of our model. By generating both semantic representations and the corresponding surface text, our system provides a criterion of correct understanding since surface texts are paired with corresponding internal representations Our theory of rational intention supports reasoning about the successes and failures of intentional actions and the ramifications of these outcomes. Also, this theory captures relative goals, deadline goals, and recurring goals. Previous theories of intention did not adequately handle as broad a class of goals. Finally, our theory of intention supports the conclusion that an agent which adopts a goal will commit to actions in pursuit of that goal. In addition to its application to narrative understanding, our theory of rational intention is applicable to multi-agent planning and speech-act processing
acase@tulane.edu