Language change and non-fictional texts – a large-scale investigation of Late Modern Swedish (1800–1950)

This project aims to investigate language change in written Late Modern Swedish, with a specific focus on non-fiction.

We will investigate changes over time and how language innovations spread across different types of texts and contribute to the establishment of language norms.

To conduct this study on a large scale, we need to use computational linguistics methods, specifically tools for automatic morphosyntactic analysis.

However, the analysis of older texts poses several challenges, such as poor OCR quality, variations in spelling and writing norms, and a lack of available annotated data for training and evaluation.

To tackle these issues, we will create new datasets with different levels of annotation, which will enable us to conduct a study on how to effectively evaluate tools for older Swedish, and we will explore methods to adapt tools from modern to older Swedish and to post-correct OCR errors.

The project will be carried out as a cross-disciplinary PhD project, supervised by two senior researchers from computational linguistics and Swedish language and linguistics, respectively.

We will perform a large-scale analysis of changes to the Swedish language from 1800 to 1950, shedding light on linguistic change as well as language norms.

In addition, the project will contribute annotated datasets, models for consistent morphosyntactic analysis, and best practices for evaluating historical Swedish, which will be valuable for future research on historical Swedish.