Quantifying domain-specificity: the occurrence of financial terms in a general corpus
Gisle Andersen · BIBSYS Brage (BIBSYS (Norway)) · 2008
In this paper, I investigate the transfer of terminology from professional discourse to language in general from a corpus-linguistic point of view. Domain-specificity can be approached from a qualitative and a quantitative angle, and the current paper has its main focus on the former approach. Frequency of use outside the specialist field can be seen as a pivotal indicator of terminology transfer. This factor is a reflection of other underlying factors, such as commercial value and public interest. The paper introduces corpus-based methods for the empirical study of transfer of terminology and presents preliminary results from the domain of business and finance. Specifically, the inventory of the KB-N termbase and that of the Norwegian newspaper corpus (NNC) are compared. It is shown that a common vocabulary can, but need not entail a conceptual overlap of the two datasets. Furthermore, the study focuses on differences between multiword and monoword terms, showing that the former type represents a high degree of ambiguity while the latter type represents meanings more specific to the domain. Moreover, it is shown that low-frequency items common in KB-N and NNC are generally a more reliable indicator of conceptual overlap between the two datasets. Finally, as can be expected, the occurrence of KB-N terminology is considerably higher in the business and finance newspaper than in the other parts of the newspaper corpus.