The linguistic affordances of campaign context: Using discourse analysis, corpus linguistics and network analysis to measure narrative cohesion and issue preference
This paper presents a multi-methods ‘lexical fields’ approach to the analysis of how political campaigns vary in terms of their linguistic affordances. Communication as action in the public sphere takes place within a complex context of associated meanings that structure the affordances of an issue sector, encouraging some ways of speaking and discouraging others. Genre expectations, ideological perspective, and cultural meanings all influence what we say, resulting in semi-stable patterns of representation around important events, issues and identities. I show how network analysis can be combined with corpus linguistics and discourse analysis to explore the online semantic profile of three campaigns around homelessness in a major metropolitan city. Rather than relying only on computational tools, this methodological combination allows for the iterative construction and measurement of those lexical fields that represent important areas of cultural meaning for a research question. As stable yet contested aspects of our society, these lexical fields can be seen as battlegrounds of associated meanings – such as whether people labeled as ‘homeless’ are dangerous, dirty drug addicts. The analysis suggests that not all issues are equally open to contestation, with the three issues analyzed here displaying widely differing levels of narrative cohesion as measured by network centralization metrics. Further, each issue area relied on a different set of lexical fields to tie the network together, privileging associations between criminality and policing in one issue network compared to shelter and home in another. Methodological and phenomenological implications are discussed.
Toft, A. (2014, February 21st). The linguistic affordances of campaign context: Using discourse analysis, corpus linguistics and network analysis to measure narrative cohesion and issue preference. Paper presented at the SUNBELT conference for the International Network for Social Network Analysis,, Saint Petersberg Florida.