Applying corpus and computational methods to loanword research : new approaches to Anglicisms in Spanish

Understanding both the linguistic and social roles of loanwords is becoming more relevant as globalization has brought loanwords into new settings, often previously viewed as monolingual. Their occurrence has the potential to impact speech communities, in that they have the capacity to alter the semantic relationships and social values ascribed to individual elements within the existing lexicon. In order to identify broad patterns, we must turn towards large and varied sources of data, specifically corpora. This dissertation aims to tackle some of the practical issues involved in the use of corpora, while addressing two conceptual issues in the field of loanword research – the social distribution and semantic nature of loanwords. In this dissertation, I propose two methods, adapted from advances in computational linguistics, which will contribute to two different stages of loanword research: processing corpora to find tokens of interest and semantically analyzing tokens of interest. These methods will be employed in two case studies. The first seeks to explore the social stratification of loanwords in Argentine Spanish. The second measures the semantic specificity of loanwords relative to their native equivalents.