BibTeX

@article{broden-etal-2026-fran-362485,
	title        = {Från ordfrekvenser till generativa modeller: Metodologisk reflektion kring datadrivna analyser av litteraturkritik},
	abstract     = {In this article, we discuss the possibilities and limitations of three different data-driven methods for replicating an earlier study in literary scholarship, Lina Samuelsson’s Kritikens ordning: Svenska bokrecensioner 1906, 1956 och 2006 (2013). We employ methods that can be said to correspond to three phases of language technology development in order to analyse research material selected according to the same principles as in Samuelsson’s study, but on a larger scale. For copyright reasons, these experiments were carried out on our material from the early 20th century.

Drawing on the Russian Formalist Viktor Shklovsky’s concept of “defamiliarisation”, we emphasise the potential of data-driven methods to create a distancing effect that prompts different perspectives on familiar research material. Beginning with word frequency analysis, we show how a simple quantitative method can contribute interesting yet methodologically rigid perspectives on a collection of book reviews. We proceed by describing how topic modelling produces results that appear analytically flat but can nonetheless sharpen the analytical gaze on the material. Finally, we show how generative language models can contribute with substantial analytical perspectives on patterns of evaluation in literary criticism, while at the same time introducing a distinct level of methodological complexity. In conclusion, we suggest that the use of data-driven methods in literary studies calls for critical reflection not only on digital and traditional methods individually, but also on the interplay between them.

Finally, we show how generative language models can contribute with substantial analytical perspectives on patterns of evaluation in literary criticism, while at the same time introducing a distinct level of methodological complexity. In conclusion, we suggest that the use of data-driven methods in literary studies calls for critical reflection not only on digital and traditional methods individually, but also on the interplay between them.},
	journal      = {Tidskrift för litteraturvetenskap},
	author       = {Brodén, Daniel and Samuelsson, Lina and Alfter, David and Karimi, Aram},
	year         = {2026},
	volume       = {56},
	number       = {3},
	pages        = {21},
}