Corpus Approaches to Analysing Uncertainty and Ignorance in Academic Discourse

The article provides an overview of corpus approaches to researching linguistic practices for dealing with ignorance and uncertainty. Uncertainty and ignorance are first and foremost epistemological or socio-psychological categories rather than linguistic ones. But they can be applied to a corpus linguistic setting. Based on a presentation of the central terms and their relevance for digital corpus research, this paper exemplifies proposals for operationalisations using a corpus of political science texts from the field of International Relations (DIReC). It gives an overview of various methods of researching ignorance and uncertainty in academic discourse, focusing on lexicon-based, annotation-based and pattern-search-based approaches as well as combinations thereof. The structure of the explanations reflects a central conflict of aims: On the one hand, corpus-based research on ignorance and uncertainty requires a precise, interpretive approach to the contextual meaning and epistemic function of each individual piece of evidence. On the other hand, it seems advantageous to investigate the largest possible corpora for reasons of reliability. The final section presents an application sketch that addresses and exemplifies several methodological problems. It compares uncertainty markers in political science discourse as found in the DIReC corpus with those in conspiracy theories, drawing on the LOCO corpus and journalistic discourse represented in a reference corpus of US newspapers.