Semantic prosody describes the tendency of words to occur in evaluative contexts. Certain words possess clear positive or negative meanings, such as success and failure. Other words have less independent evaluative meanings, but tend to occur in positive or negative environments, such as cause. This thesis studies the boundary between these two types of meaning, focusing on how separate senses of words affect their semantic prosodies, and asking why these prosodies develop and why they vary across different types of text.
When words appear in the same kinds of contexts over time, they can be coloured by their environment, creating an expectation that they belong there. For example, we are more likely to assume that something is negative when it is caused than when it is produced. Because of this, exceptions to the trend, such as cause success and produce failure, stand out and can be used deliberately for rhetorical effect. In addition to time, factors such as separate senses and text types also affect semantic prosody. For example, fabricate conveys a different meaning in relation to evidence than it does to steel. Further, these separate senses are likely to appear in different types of texts.
This thesis demonstrates that the development of semantic prosodies is linked with the emergence of new senses, that separate senses are associated with different evaluative contexts, and that variation across text types can be attributed to an unequal distribution of senses.
Mathias Skogly Russnes successfully defended his thesis on 6 May 2026.
Trial lecture
Designated topic: Corpus approaches to evaluation.
Evaluation committee
- Dr. Svetlana Vetchinnikova, University of Helsinki (first opponent)
- Dr. Gavin Brookes, Lancaster University (second opponent)
- Professor Hilde Hasselgård, University of Oslo (committee administrator)
Chair of the defence
Supervisors
- Professor Signe Oksefjell Ebeling, University of Oslo
- Dr. Nicholas Groom, University of Birmingham