Open Access

Lesley-Ann Kern, Dorothee Meer

• Narratives (defined as a discursive structure based on local and temporal indications and actors as part of a development of action helping to solve problems) are similar to linguistic phenomena such as metaphors or argumentative patterns usually analyzed in qualitative linguistic research. In this paper, our idea is to combine these qualitative linguistic methods with quantitative corpus linguistic approaches such as named entity recognition (NER). We present a pilot study in which we use current NER technology to semi-automatically detect the narrative of energy partnership in media discourse. Based on a manual pre-study of a part of our corpus, we defined detectable expressions with a semantic relation to the subject that we link to occurrences of the narrative. Our study shows that the automatic detection of narratives still proves to be difficult, yet using NER we could identify pre-defined narratives in our corpus. This shows that quantitative approaches can support qualitative text analysis and should therefore be considered when working with digital corpora to gain extensive insights into the material.