Lab

LexiCon Research Group

About the lab

The LexiCon Research Group at the University of Granada has been doing research in Terminology, Lexical Semantics, Lexicography, Metaphor, Translation, and Specialized Knowledge Representation since 1994. Over the last decade, the members of LexiCon have focused the design of specialized multilingual knowledge bases and ontologies, such as Oncoterm and EcoLexicon, which can be used as terminological and translation resources. These knowledge bases have been created using on a robust model of semantic analysis, based on the creation of lexical templates derived from corpus and dictionary analysis. The approach applied is known as Frame-Based Terminology. This application of this model has led to a wide range of publications.

Featured projects (1)

Project
The main objective of CONTENT is to fully exploit the contents and components of EcoLexicon for purposes of translation and natural language processing. Accordingly, this project will create and implement a prototype for the terminology-enhanced translation of specialized environmental texts. This means expanding the architecture of the relational database where EcoLexicon is stored, as well as enriching the following modules: (i) the linguistic module (inclusion of relations between terms); (ii) the conceptual module (specification of non-hierarchical relations based on paradigms encoded in the phraseological module); (iii) phraseological module (expansion of syntagmatic relations enhanced with different types of collocational information). This prototype will make EcoLexicon’s data available online to users in contexts that provide a selection of semantic, syntactic, and pragmatic information specifically related to the terms in the source text. The enhancement of the modules in EcoLexicon as well as the design of the prototype require the use of more effective terminographic methods and the extensive semi-automatic processing of the corpus based on the extraction of knowledge patterns. Parallel to the design and implementation of the prototype, still another goal is to facilitate the interoperability of EcoLexicon by linking it to other resources by means of Linked Data, a technology that publishes structured data and links them to information in other resources in compliance with Semantic Web standards. The data in the formal ontology version of EcoLexicon will thus be accessible by means of SPARQL queries. Furthermore, Ecolexicon data will be limited to the information in GEMET and AGROVOC or DBpedia with a view to offering an open resource integrated in the Semantic Web or more concretely, in Linguistic Open Data. The linking process will be semi-automatically performed by means of RDF properties (e. g. rdf: SeeAlso), OWL (e. g. owl:SameAs) and SKOS (e. g. skos:broader). In this way, the conceptual and linguistic information in Ecolexicon can be transformed into a disambiguation resource. At the same time, the content of the linked resources will be exploited in the terminology-enhanced prototype for assisted specialized translation. Finally, the inventory of semantic relations in Ecolexicon and its underlying conceptual structure will be validated by an fMRI experimental study, based on the successful results of a previous pilot study (Faber et al. 2014). The objectives will focus on the representation, storage, and processing of specialized concepts as well as their semantic relations. Following Muelhaus et al. (2014), the subjects (experts and non-experts) will be subjected to different stimuli (images, terminological designations, and terms associated with different types of relation) in order to analyze which type of semantic relation most facilitates the comprehension of a concept and whether vertical and horizontal semantic relations have different brain activation patterns.

Featured research (6)

This chapter develops the theoretical premises and the semantic, syntactic, and pragmatic microtheories underlying Frame-Based Terminology and the organization of specialized knowledge. It also explains how these premises have been implemented to formulate terminological definitions, model specialized knowledge concepts, create semantic networks, and design specialized knowledge resources. Conceptual modeling in knowledge resource design should capture both the micro- and macrocontexts of concepts. This not only means structuring individual term entries but also capturing the relationships between them. These relations reveal the most frequent combinations and activations of specialized knowledge units, which are indicative of some type of large-scale knowledge structure in which individual concepts are meaningfully related to each other.
Resumen: Este trabajo examina algunos proyectos que ofrecen rutas de acceso al léxico innovadoras y que constituyen un primer paso hacia un diccionario más inteligente. Estos incluyen los diccionarios de colocaciones; aquellos que especifican las restricciones semánticas y léxicas entre un predicado y sus argumentos; aquellos que describen las propiedades lingüísticas de una entrada léxica en virtud de un marco o marcos; incluso aquellos que proporcionan una organización conceptual del definiens, y no del definiendum. Sin duda alguna, la transformación digital, la Inteligencia Artificial, la minería de datos y textos y el aprendizaje automático de máquinas han abierto un nuevo abanico de posibilidades para el diseño de recursos lexicográficos que debería redundar en recursos más inteligentes y conectados. Se concluye el trabajo planteando algunas ideas, a modo de rasgos definitorios y con un carácter muy prospectivo, sobre lo que debería ser un diccionario digital inteligente o un diccionario 5.0. Palabras clave: relaciones léxicas, colocaciones, ontologías, definiciones mínimas, datos lingüísticos enlazados, minería de datos. Abstract: This paper examines a set of lexicographic projects with innovative routes of access to lexical knowledge, and which constitute a first step towards a more intelligent dictionary, These projects include: (1) collocation dictionaries that specify the relations between a base word and its collocate; (2) dictionaries that make explicit the semantic and lexical restrictions between a predicate and its arguments; (3) lexical resources that describe the linguistic properties of a lexical entry within the context of its frame or frames of activation; (4) dictionaries that provide a conceptual organization of the definiens, instead of the definiendum. Without a doubt, the Digital Era (artificial intelligence, data and text mining, and machine learning) has opened up a vast range of possibilities, which will lead to intelligent lexicographic resources that are more intelligent and interconnected. This chapter concludes with some ideas and proposals about the characteristics of a dictionary 5.0 of the future. El principal objetivo de un diccionario es representar el conocimiento relacionado con un grupo de palabras y adecuar la cantidad y el formato de estos datos a las necesidades de los usuarios finales, que podrán utilizar esta información bien para codificar o para descodificar un mensaje. Además, el valor de un diccionario reside no solo en el poder expresivo de sus definiciones y las categorías de datos de su microestructura, sino en su configuración global y en su organización, es decir, en establecer qué rutas de acceso ofrece al usuario para acceder a estos datos. En este contexto, un diccionario se concibe como un repositorio dinámico y flexible donde reside una intrincada red de relaciones entre palabras, y tal es su formato que se suele afirmar que cuanto más se asemeje al lexicón mental, más valioso resultará como recurso. Precisamente, uno de los aspectos más importantes del lexicón mental es el modo en que se representan e interrelacionan las palabras, de modo que sea posible acceder a ellas de forma más eficiente. Esta especificación de la visión más amplia es lo que la mente hace con tanta facilidad, y lo que los recursos léxicos sólo pueden lograr con un gran esfuerzo. Como veremos más adelante, incluso en diccionarios en los que las redes léxicas y/o relaciones semánticas están codificadas, únicamente existe una similitud parcial entre estas representaciones y To cite this article: Mairal-Usón, R., Faber, P. (2021). "Routes of access to the lexicon in a lexicographic environment". Revista de Lingüística y Lenguas Aplicadas, 16, 63-79. https://doi.
A well-designed terminological knowledge base is a structured repository of linguistic data, which is enriched with metadata and structured according to specific classification schemes and concept-based analysis. There is now no limit to the quantity and type of information in each entry since the digital age has liberated lexicographical and terminological resources from space constraints. More specifically, a terminological knowledge resource for environmental translators should be tailored to the specific needs of users who work with this multi-faceted type of specialized text. Accordingly, the type and configuration of the environmental information should reflect the micro and macrostructural design of the resource, and provide frame-like structures in which concepts and terms are dynamically related to others. Designers must also decide how each type should be accessed, and in what sequence. The map of the resource can have an underlying conceptual frame that allows users to derive the maximum benefit from it. This chapter provides an overview of how the needs of environmental translators and explains how they can be met in the design of a knowledge resource. This is illustrated by the type and number of data fields, their underlying principles, and mode of visualization.
EcoLexicon es una base de conocimiento terminológica multilingüe sobre ciencias medioambientales desarrollada desde 2003 por el grupo de investigación LexiCon de la Universidad de Granada (España) y constituye la aplicación práctica de la teoría de la terminología basada en marcos. El presente artículo describe el funcionamiento de EcoLexicon y presenta sus últimos avances, que incluyen un nuevo corpus y una gramática semántica de word sketches en inglés, una reforma del módulo fraseológico, un enfoque flexible a las definiciones terminológicas y la representación conceptual mediante imágenes.
Contemporary research has focused on how concepts are represented and organized in the mind, leading to neurocognitive theories such as grounded cognition or embodied cognition. These theories have greatly influenced further studies in linguistics and terminology. In this way, conceptualization, categorization, and knowledge organization are the foundation of cognitive-oriented terminology theories which highlight the relevance of situated knowledge structures, such as Frame-based Terminology. Accordingly, the practical application of Frame-based Terminology is EcoLexicon, a dynamic terminological knowledge base on environmental science. Concepts in this terminological resource are domain-specific within the Environmental Event, a model that interrelates concepts by assigning them different roles. However, the Environmental Event does not include specific category types to annotate these concepts ontologically. Therefore, this paper presents a process of ontological knowledge enhancement in EcoLexicon. This process was mainly based on the categorization of its concepts in semantic classes with a multidimensional approach. As a result, EcoLexicon was ontologically enhanced not only in terms of this categorization, but also through a redesign of the conceptual categories module, which involved modifying the existing category hierarchy and implementing new features focused on describing the combinatorial potential of concepts and categories (i.e. the conceptual combinations function and the ontological view).

Lab head

Pamela Faber
Department
  • Faculty of Translation and Interpretation

Members (17)

Pilar León Araúz
  • University of Granada
Clara Inés Lopez-Rodriguez
  • University of Granada
Maribel Tercedor
  • University of Granada
Miriam Buendía Castro
  • University of Granada
Juan Antonio Prieto Velasco
  • Universidad Pablo de Olavide
Arianne Reimerink
  • University of Granada
Antonio San Martín
  • Université du Québec à Trois-Rivières
Melania Cabezas-García
  • University of Granada