Corpus-based genre analysis of online music reviews: Exploring the rhetorical structure

Rhetorical move-step analysis has originally been applied to academic texts (Swales 1990). Recent years, however, have seen this approach being employed to both discourse and genre analysis of other written genres. Reviews appear to be an interesting instance of texts as they have a transformative impact on the reality, i.e. they may significantly influence the success or failure of a particular album or the public’s and the experts’ perception thereof. While book and film reviews have been subject to genre analysis (Motta-Roth 1995, De Jong and Burgers 2013), music reviews have received scant scholarly attention in this respect. In order to fill this gap, this paper aims to provide a move-step genre analysis of a specialised corpus of 75 online music reviews from three online reviewing websites, namely Pitchfork, NME and slantmagazine. Three subcorpora of music reviews, each consisting of 25 texts published between 2021 and 2023 were compiled for the purpose of this study (Koester 2010). This paper investigated the following research questions: What differences in rhetorical structure can be established according to the positive or negative evaluation in the review? What differences in rhetorical structure can be established according to the website? What lexical bundles are used by reviewers to evaluate artists and their music? The corpus was manually annotated with the use of a rhetorical function codebook including steps and moves developed on the basis of similar studies employing structural move analysis (Moreno and Swales 2018, Bhatia 1993) and adjusted to the characteristics of the analysed corpus. A pilot study was conducted during which a sample of reviews was coded by two annotators. After achieving a high level of intercoder agreement, final codebook was created. Respective parts of reviews were assigned a particular rhetorical step, e.g., defining the theme of the album. Steps were then grouped into moves, e.g., summary. The same steps were extracted from each subcorpus to examine what patterns are most frequently employed to realise given steps. Differences between American (Pitchfork, slantmagazine) and British (NME) reviews were also deliberated on to shed light on rhetorical differences in these two English varieties. It was also established what lexical bundles are most frequently employed to realise particular rhetorical moves and steps. Preliminary results indicate that the step dimension of the analysis is the main feature differentiating the three platforms from one another.