“Steps” towards a corpus of SCOTUS opinions annotated using a Swalesian approach

Abstract

In the tradition of Moreno and Swales (2018), this paper presents the creation of a manually annotated resource for supporting teaching English for Legal Purposes (ELP) and for Natural Language Processing (NLP) purposes. After justifying the use of Supreme Court of the United States opinions, we define our coding scheme by adapting the move model of rhetorical structure in specialized discourse. We describe the methodology and the implementation of the annotation campaign. We analyze how our methodology and the resulting annotation scheme diverge from those described in the literature as well as the advantages that these divergences afford. In addition to the research article, we release several supplementary materials which aim to make the process transparent and serve other researchers aiming to annotate specialized discourse with the help of machine learning techniques.