Pragmatic Annotation for a Multi-Layered Analysis of Speech Acts: A Methodological Proposal

Abstract

Within the recently-coined sub-field of corpus pragmatics, one of the areas of interest is the study of speech acts and, specifically, how it can profit from the adoption of this methodological approach. However, the acknowledged lack of correspondence between speech acts and linguistic forms makes basic form-based corpus searches unreliable in retrieving speech acts from a corpus. In fact, function-toform corpus research can prove much more fruitful in carrying out this kind of study, but it usually requires time-consuming manual annotation, which in turn means that there have been few attempts to employ this methodology. As a contribution in this new direction, this study will showcase a function-toform approach to investigating speech acts of agreement and disagreement in spoken Catalan. Through this example, this paper aims to show the benefits of designing, compiling, transcribing and, especially, annotating one's own corpus for the study of speech acts. In order to annotate data for the study of speech acts, a complex and multi-layered annotation system was designed and manually applied, so that all the different aspects that play a relevant role in the expression of agreement and disagreement could be covered. In addition to discussing the findings from this study, it is argued that the possibilities of exploitation provided by the resulting annotated corpus far outweigh the time cost and open the door to in-depth analyses of speech acts and politeness in naturally occurring spoken data. function-to-form approach to the study of speech acts and central to this is the careful design of the corpus and annotation system (as detailed below). Corpus design for the study of agreement/disagreement According to Kasper and Dahl (1991), research instruments are one of the main causes of variation in pragmatic studies, together with sociolinguistic elements. Discourse completion tests (DCTs) and role plays have amassed a long tradition in pragmatics, but the combination of pragmatics with corpus linguistics turns the use of naturally-occurring discourse into a basic requirement. Once a corpus is compiled or obtained, corpus linguistics can be brought in as a methodological tool (O'Keeffe and Walsh 2012: 161), in order to lead to "greater depth of analysis in combination with another theoretical framework". The quantitative nature of corpus linguistics can be combined with different frameworks which focus on a qualitative approach, such as conversation analysis (e.g.

Pragmatic annotation of a domain-restricted English-Spanish comparable corpus

“…Besides, most pragmatic studies are relatively small-scale qualitative analyses concentrating on spoken language data samples (Archer et al, 2008: 613;Milá-García, 2018). Pragmatically annotated corpora of written texts are still rare but see Marín-Arrese's CESJD tagset (2017,2019) or Weisser's in-progress TART dataset proposal (2018: 280ff).…”