A theoretical approach to automatic loanword detection
Marisa Köllner · Universitätsbibliothek Tübingen · 2013
For several years, computational methods found their way into humanities. Especially in the field of computational linguistics several analysis andmethods are studied. It is not surprising that computational analysis arouse interest in the field of historical linguistics. Due to such methods, language evolution can be studied from another point of view. Biological and linguistic evolution show certain parallels. Especially the parallels between phylogenetics and linguistics arouse the interest of combining both fields. Phylogenetics provide a great number of mathematical and computational methods for computing di erent tasks. Based on the parallels, the methods can be adapted into historical linguistics. In historical linguistics, the process of borrowing is a well-known evolutionary process where words are borrowed from one language and adapted into another. Borrowing has its corresponding parallel within phylogenetics, namely horizontal gene transfer. Horizontal gene transfer is the process of transferring genes from one organism to another. The similarity between borrowing and horizontal gene transfer is the transfer of genes or words whereas the organisms or languages are not related. Phylogenetics provides several computational methods and analysis to detect horizontal gene transfer. The methods might be adapted into linguistics to detect borrowing. This paper introduces the background of borrowing and phylogenetics as well as the combination of both fields. The new tree-based approach should indicate if provided methods of phylogenetics can be adapted into linguistics for the detection of borrowing.