Cassandra: Explaining and predicting short-term language change in Modern Swedish
The projekt Cassandra is a collaboration of researchers working at the Department of Swedish, Multilingualism, Language Technology (principal investigator Aleksandrs Berdicevskis, together with Yvonne Adesam och Nina Tahmasebi) and the Department of Languages and Literatures (Evie Coussé). The central question in the project is: Is it possible to predict linguistic changes?
Background and purpose
Human languages are constantly changing. But what makes the language change? By explaining the recent changes, the researchers also want to see if it is possible to predict future changes in the Swedish language.
Some examples of changes in Swedish during the 20th and 21st centuries are that verb forms in the plural, such as gingo (went), have disappeared; that the word grym (cruel) has developed a new positive meaning, which is almost a direct opposite to the original meaning, and is becoming dominant; that people more and more often say starkare än mig (stronger than me) and not starkare än mig (stronger than I ); that the genderneutral pronoun hen (he/she), which was first proposed in 1966, has in fact begun to be used in recent years.
Linguists are constantly trying to explain changes. The disappearance of gingo (went) can be explained by the fact that language users like to simplify, at least under certain conditions. Similar simplifications have taken place in many other Germanic languages, including English. Negative words like grym (cruel) often undergo similar changes of meaning (from a negative meaning to a reinforcing word with a positive meaning). It can be assumed that hen (he/she) began to be used as a result of social changes (i.e. the struggle for equality).
Cassandra want to evaluate whether explanatory theories of language change have the ability to predict change. The question has received little attention in linguistics but needs to be answered if one wants to estimate how reliable the existing explanations are. By analyzing incorrect predictions, we will examine whether the theories are wrong and, if so, what they do not take into account.
For the quantitative survey, the project will use large amounts of text with linguistic analysis, so-called corpora, which contain posts from social media over the past twenty years. They will calculate diachronic developmental paths for different phenomena: frequencies of words, word forms, word meanings and expressions. The predictions will be based on existing hypotheses about the mechanisms of language change and use both linguistic and non-linguistic information (mainly the social connections between forum users).
Goal and results
The project will provide theoretical results, but also new language resources (the corpora enriched with information on the structure of social networks and on language changes) as well as new methods, relevant to anyone interested in how changes spread through society.