Since antiquity figures have been widely studied within the framework of rhetoric, although contemporary rhetoric has tended to disregard their importance and relegate their study to the domain of literary criticism. However, since the 1980s, there has been a renewed interest in figurative language not only in literary studies, but also in other fields of research. Indeed, research on figuration has emerged as a new and distinct discipline, namely figurative language studies. However, within this framework, metaphor and irony have received the greatest attention, while other non-literal forms have been largely ignored. This is certainly the case of hyperbole, a long neglected trope despite its ubiquity in discourse. The present study aims to make a contribution to the literature on exaggeration and so by extension to figurative language theories in general. Not all aspects of figuration have attracted equal interest among researchers. With a few exceptions, most attention has been directed at explaining how figures are comprehended, given their non-literal nature. In contrast to understanding, the question of production has been largely ignored. Similarly, although the reception process, in terms of understanding, has been widely studied, almost no attention has been devoted to listeners responses to figures and their collaboration in a joint construction of non-literal frames. Rather than addressing comprehension, this study concentrates on the production process and usage of exaggeration, since these fields of research have been largely ignored in the literature on the subject. It aims to provide a general framework for the description and understanding of hyperbole in interaction, mainly from a production viewpoint but without totally disregarding the reception process, since special attention is devoted to the interactive dimension of exaggeration. This aim is formulated in terms of the following objectives: Objective 1: to provide an adequate definition of the notion of hyperbole; to list the criteria for identifying the trope so that non-exaggerated uses of expressions can be excluded. Objective 2: to set up a classification of overstated items according to semantic field, grammatical category, auxesis or meiosis, and interactivity with other figures. Objective 3: to explore the long neglected production process of hyperbole, both in terms of usage and functions. Objective 4: to examine the interactive nature of the trope, as an activity collaboratively constructed between speaker and hearer, by studying listeners reactions and their own further contributions to overstatement. As for the theoretical framework, this study combines pragmatic and conversational-analytical methods with a corpus-based approach to the study of hyperbole. Exaggeration is a purely pragmatic phenomenon since it is entirely dependent on context. On the other hand, since overstatements are not one-off but complex lexico-grammatical items, they need to be examined within the constraints of placement, sequencing and turn-taking of conversational analysis. Finally, a major benefit of corpus-based research, only recently applied to the study of figures, is that it grounds its theorising on empirical observation rather than linguistic intuition. Besides, the use of corpora grants certain benefits, such as the use of naturalistic data, automatic access to context, evidence of interactivity and hyperbolic cues, wide coverage of genres, etc. The data examined consists of naturalistic spoken texts, totalling 52,000 words, extracted from the British National Corpus. The focus is on oral discourse, since not a great amount of research exists into everyday spoken hyperbole. The bulk of research has been conducted in written language or relies on artificial and elicited data. The aim was to demonstrate that although traditionally relegated to the literary text, hyperboles are rather ubiquitous strategies in everyday language. This idea adheres to a prevailing view among figurative language researchers, namely that cognition is partly figurative. __________________________________________________________________________________________________ RESUMEN El lenguaje figurado ha sido, desde antiguo, ampliamente estudiado en el marco de la retórica. Más recientemente, las figuras retóricas parecen haber despertado un nuevo interés en otras disciplinas. De hecho, dichos estudios están empezando ya a ser considerados como una disciplina en sí: las teorías del lenguaje figurado. Dentro de este marco, metáfora e ironía han sido ampliamente estudiadas y como consecuencia de su estudio intensivo, otras figuras han sido marginadas. Éste es el caso de la hipérbole, figura retórica largo tiempo olvidada a pesar de su ubicuidad y tradicionalmente relegada al dominio de la crítica literaria. El presente estudio, en lugar de examinar el proceso de comprensión, se centra en la producción y uso de la exageración, temas escasamente tratados en la literatura existente. Intenta proporcionar un marco general para la descripción y comprensión de la hipérbole en la interacción, principalmente desde el punto de vista de la producción, pero sin descartar del todo el proceso de recepción, dado que su dimensión interactiva como figura creada conjuntamente entre hablante y oyente cobra especial relevancia. Dicho propósito se materializa en los siguientes objetivos: Objetivo 1: proporcionar una definición adecuada de la noción de exageración que nos permita diferenciarla de figuras afines y excluir los usos no exagerados de las expresiones. Objetivo 2: clasificar los elementos hiperbólicos de acuerdo con los siguientes parámetros: campo semántico, categoría gramatical, auxesis o meiosis, e interacción con otras figuras. Objetivo 3: explorar el descuidado proceso de producción de la figura en términos de uso y función. Objetivo 4: examinar la naturaleza interactiva del tropo, como actividad conjunta entre hablante y oyente, mediante el estudio de las respuestas y demás contribuciones del receptor al acto hiperbólico del hablante. Este estudio fusiona tres tipos de enfoque: pragmática, análisis conversacional y lingüística de corpus, a la hora de estudiar la exageración en el discurso oral, dado que son escasos los estudios de esta figura en el habla cotidiana. El corpus examinado abarca unas 52.000 palabras y es una selección de textos orales extraídos del British National Corpus.