Material addicional: http://hdl.handle.net/10230/6323 / Aquesta tesi doctoral té com a objectiu principal establir les bases per a un sistema de detecció semiautomàtica de relacions conceptuals a partir de textos especialitzats. Per arribar a aquest objectiu, la tesi presenta una definició de relació conceptual en el marc d'un enfocament comunicatiu de la terminologia i, a partir d'aquesta definició, l'autora detecta i aplica diversos marcadors lingüístics verbals per aïllar fragments textuals que continguin unitats de coneixement especialitzat, vehiculades mitjançant termes i relacions conceptuals. D'aquesta manera es pretén recuperar fragments de coneixement especialitzat a partir dels nusos de coneixement i de les seves relacions en textos de l'àmbit del genoma humà.Una de les aportacions principals d'aquesta tesi és l'aplicació d'una tipologia de relacions conceptuals validada empíricament a partir de textos especialitzats a la creació i alimentació d'una ontologia sobre el genoma humà. I orientat a complir l'objectiu general d'aquest treball, la detecció semiautomàtica de relacions conceptuals, l'autora proposa estratègies sintàctiques i semàntiques que permetin refinar al màxim aquests elements clau en l'organització de la informació especialitzada, combinant aquestes estratègies amb un detector i un extractor de terminologia. / The main goal of this Ph. dissertation is to establish the ground basis for a semiautomatic conceptual relations detector on the basis of specialised texts. In order to attain this goal, the Ph. dissertation includes a new definition of conceptual relations in a communicative approach to terminology. From this definition, the author detects and applies different linguistic verbal markers to isolate textual fragments containing specialised knowledge units expressed by terms and conceptual relations. The aim is to retrieve specialised knowledge fragments from knowledge nodes and the relations they establish among them in the human genome domain.One of the main contributions of this work is the application of a typology of conceptual relations empirically validated on the basis of specialised texts on the construction and updating of an ontology about the human genome domain. As for the general goal, the semiautomatic detection of conceptual relations, the author proposes syntactic and semantic strategies for the maximum refinement of the detection of these key elements in the specialised information organisation, together with the use of a term detector and extractor.Podeu consultar material addicional a http://repositori.upf.edu/handle/10230/6323

The conceptual relation theory postulates that English noun-noun compound words (e.g., snowman) have an underlying predicate structure that is not present in the surface form, but is recovered during compound processing (e.g., man made of snow). The relational nature of constituent binding in compound words marks them as a linguistic construction that is distinct from both the simplex words (monomorphemic) and other complex words (derived and inflected words) previously examined in the context of verbal working memory. In short-term memory research, a growing body of evidence suggests that semantic properties of words influence verbal recall; however, such effects have not been examined in the context of compound conceptual relations. The present study investigated the possible effects of compound conceptual relations in verbal working memory via an immediate serial recall task. The task was designed to examine whether sharing of an individual relation leads to facilitative or inhibitory effects for compounds associated with that relation and, more generally, whether this semantic property of compound words contributes to their recollection from short-term memory. Evidence from the serial recall experiment suggested an effect of compound relation priming in working memory. Relational similarity between recall list items appeared to inhibit recall performance. The thesis discusses how this may be the result of increased competition between compound constituents as a result of heightened constituent-level activation during word recall. This effect was not observed in relations that appeared to be overly general, suggesting that the effect is only present when compound words are matched according to salient, sufficiently specified relations. / Thesis / Master of Science (MSc)

El objetivo de esta tesis, es reunir información gramatical suficiente que permita determinar qué características deben reunir las bases verbales del español para ser capaces de originar un adjetivo deverbal de participio (ADP), y, basados en ello, poder describir cómo opera el proceso de activación de valor especializado en los términos N+ADP del dominio de la economía. Estas construcciones sintácticas mínimas son muy productivas en los discursos de ámbitos especializados, sin embargo la naturaleza híbrida de la forma participial acarrea muchos conflictos a la tarea de Procesamiento de Lenguaje Natural (PNL). Esta aproximación al análisis de los ADP es lingüística, está anclada teóricamente en la Teoría Comunicativa de la Terminología (TCT) e intenta ser el punto de contacto entre teoría y aplicación. / The goal of this thesis is to pinpoint the grammatical information that is necessary to determine which Spanish verb stems give rise to an adjectival participle (AP). This information will allow us to describe the linguistic indicators that, within the domain of economy, activate a specialized meaning in those terms that have the structure AP+noun. These syntactic minimal constructions are highly productive in specialized discourse. Nevertheless, the hybrid nature of the participial form invokes many conflicts in Natural Language Processing (NLP) applications. This descriptive approach to the adjectival participles is linguistic in nature, based on the Communicative Theory of Terminology (CTT), intends to be the point of contact between theory and application.

L’objet de notre recherche sont les relations conceptuelles exprimées dans les définitions des dictionnaires et celles exprimées dans un corpus spécialisé. Nous avons pour but d’analyser et comparer ces relations pour identifier les relations les plus communes d’un domaine spécialisé et déterminer où ces relations se trouvent plus fréquemment. Notre approche considère que ces relations se trouvent plus souvent dans les corpus et qu’on pourrait enrichir les définitions terminologiques en incorporant ces relations conceptuelles extraites des textes spécialisés. Le domaine choisi pour cette étude est celui de la pêche durable dont nous analysons la terminologie en anglais. Les termes analysés sont extraits d’un corpus de textes de ce domaine construit pour notre étude et qui comporte des articles scientifiques et des comptes rendus d’organismes spécialisés dans le domaine de la pêche. Pour l’analyse de définitions, trois dictionnaires spécialisés en pêche ont été sélectionnés dans l’étude. L’échantillon final de termes analysés inclut 20 noms (dont 12 termes dénotent des entités et 8 termes dénotent des activités). Ces termes sont les plus spécifiques extraits du corpus avec l’extracteur TermoStat (Drouin, 2003) et définis dans au moins deux des dictionnaires choisis. Les unités lexicales du corpus sont repérées de façon semi-automatique à l’aide de la fonctionnalité word sketch, « an automatic corpus-derived summary of a word’s grammatical and collocational behavior » (Kilgarriff et al., 2010, p. 372) dans la plateforme de gestion de corpus Sketch Engine (Kilgarriff et al., 2014). Nous travaillons avec deux types de word sketches: le word sketch conventionnel fourni par défaut par Sketch Engine et l’EcoLexicon Semantic Sketch Grammar (ESSG; León Araúz & San Martín, 2018). Seules les unités lexicales les plus fréquentes sont sélectionnées de tous les résultats de l’interrogation du corpus. L’analyse des définitions se penche sur toutes les unités lexicales reliées directement au terme analysé. Nous utilisons des paraphrases dans les analyses pour identifier et valider les relations entre le terme analysé et chaque unité reliée. À la suite de l’identification des relations, nous compilons une liste de relations et nous faisons une comparaison entre les résultats du corpus et des définitions. La comparaison des types de relations repérées dans chaque source montre qu’il y a plus de types de relations dans le corpus que dans les définitions pour 70 % de l’échantillon de termes. Lorsque la comparaison examine séparément des termes dénotant des entités et des activités, plus de types de relations se trouvent dans le corpus que dans les définitions pour 83 % des entités et pour 50 % des activités. Les résultats montrent également que 54 % des types de relations repérées sont identifiés pour plus de termes dans le corpus que dans les dictionnaires. Par ailleurs, seulement 16,7 % des relations repérées sont identifiées pour plus de termes dans les dictionnaires que dans le corpus. La recherche a également identifié quels types de relations se trouvent plus souvent dans le corpus, dans le dictionnaire ou dans les deux sources pour le même terme. Ce constat a permis de classifier les types de relations dans trois groupes: les relations qui se trouvent la plupart du temps dans les dictionnaires, celles plus souvent présentes dans le corpus ou celles présentes dans les deux sources. / The object of our study are the conceptual relations expressed in dictionary definitions and those expressed in a specialized corpus. Our goal is to analyze and compare these relations to identify the most common relations of a specialized subject field and determine where these relations are more frequently found. Our approach considers that these relations are more often found in the corpus and that we could enrich terminological definitions if we include the conceptual relations extracted from specialized texts. The subject field chosen for this study is sustainable fisheries from which we analyze the terminology in English. The terms analyzed were extracted from a corpus of texts belonging to this subject field and built for the study. The corpus includes scientific articles and reports issued by specialized organizations in the field of fisheries. For the analysis of definitions, three specialized dictionaries were selected for the study. The final sample of terms analyzed includes 20 nouns (12 terms designating entities and 8 terms designating activities). These terms are the most specific terms extracted from our corpus using the term extractor TermoStat (Drouin, 2003) and defined in at least two of the selected dictionaries. The lexical units from the corpus were extracted semiautomatically using the function word sketch, “an automatic corpus-derived summary of a word’s grammatical and collocational behavior” (Kilgarriff et al., 2010, p. 372) in the corpus management platform Sketch Engine (Kilgarriff et al., 2014). We worked with two types of word sketches: the conventional word sketch provided by default in Sketch Engine and the Ecolexicon Semantic Sketch Grammar (ESSG; León Araúz & San Martín, 2018). Only the most frequent lexical units were selected from all the results of the corpus interrogation. The analysis of definitions included all the related lexical units directly linked to the analyzed term. Paraphrases were used in the analysis to identify and validate the relation between the analyzed terms and the related lexical units. Once all the relations were identified, a list of relation types was compiled, and a comparison was made between results from the corpus and the dictionaries. The comparison of the relation types found in each source shows that there are more relation types in the corpus than in the definitions for 70% of the sample. When the comparison focuses separately on entity and activity terms, more relation types were found in the corpus than in the definitions for 83% of entity terms and 50% of activity terms. Results also show that 54% of the relation types are associated with more terms in the corpus and only 16.7% are associated with more terms in the dictionaries. Additionally, the study identified which relation types are more often found in the corpus, in the dictionaries or in both sources. These findings allowed us to classify the relation types in three scenarios: relation types mostly found in the dictionaries, those more often found in the corpus and the group of relation types which are mostly found in both sources for each term.