Conference article

In the field of historical linguistics; large-scale corpora studies are a key component in identifying phenomena such as language variation and language change. Manually performed corpora studies are very time consuming and may obscure interesting changes in the sense that phenomena that are not being specifically searched for easily are overlooked. During the last couple of years the potential of language technology tools has been put forward in relation to historical linguistic research. This paper is based on an experiment of linking up several lexical resources in Swedish; which together reflect a vocabulary from Old Swedish to Contemporary Swedish. The link-up aims at identifying potential lexical change such as cases of grammaticalization and it may further be of use in other language technology applications. In our case study we are linking lemmas together with part-of-speech information given in each entry for all the lexical resources. This paper describes our first results; where we focus on the cases when information about word class differs in one of the resources. In future studies it is necessary and desirable to include more digitalized lexicon resources and confirm the analyses with corpora research. Still; the current result already shows some interesting cases of semantic change and grammaticalization. Changes in the content word system such as generalization and specialization of meaning are also exemplified in our data. Even though the links sometimes show errors that at first sight lead us towards a rong conclusion we believe that methods like the one used here may be very fruitful to future research to reach more efficiency in historical linguistics research.

Lexical linking; historical linguistics; language change