Corpus-based narrative analysis: Bridging lexical patterning and narrative structure in large-scale discourse

In response to the methodological challenge of scaling narrative analysis in discourse studies, this paper introduces an integrated framework that combines corpus linguistics (CL) and narrative analysis (NA) to detect narrative patterns across large textual datasets. This framework, termed Corpus-Based Narrative Analysis (CBNA), enables systematic identification and validation of abstract narrative templates (‘proto-narratives’) through corpus statistics and narrative theory. Drawing on Ricoeur’s concept of emplotment and the Discourse-Historical Approach (DHA), the method operationalises narrative structure through what we term proto-narratives: abstract, schema-like representations of recurring narrative trajectories. These are constructed using corpus-driven frequency, collocation, and concordance analyses and validated through qualitative analysis. This hybrid methodology provides a replicable procedure for uncovering dominant narrative forms and their institutional variation. We demonstrate this approach through a case study based on German political and media discourse on irregular migration, revealing dominant narrative forms—such as the migrant as a threat, victim, or burden—and how they differ across institutional contexts. Our method bridges the interpretive richness of narrative inquiry with the empirical scalability of corpus analysis. It offers a replicable, transparent model for narrative analysis at scale and provides new tools for tracing narrative convergence, divergence, and ideological framing across genres and contexts.