The collocation networks of stance phrases

Collocation networks provide important information on the semantic relatedness of linguistic features and their co-texts. Previous studies on collocations have largely been limited to the descriptive examination of particular lexical items or phrases, while the statistical analysis of collocational relationships of phrases performing similar functions remains underexplored. This study investigates the collocation networks of stance phrases in cross-disciplinary academic discourse from a quantitative perspective. We used GraphColl to identify the collocates of stance phrases in two purpose-built corpora consisting of agriculture and economics research articles, and we then generated the collocation networks of different categories of stance phrases in the two corpora. The findings show a complex and intricate collocation network which displays both the within-stance category collocations and the collocational patterns between stance phrases and other semantic domains. Significant discipline-specific features were identified in the collocation networks of stance phrases. These findings contribute to our knowledge of how different categories of stance operate with their surrounding linguistic domains to assist writers to develop convincing arguments, construct coherent texts and communicate effectively with envisaged readers.