Abstract

Sentiment analysis has received increased attention during the last decade, with a growing number of both research initiatives and industrial applications. Accordingly, methodological approaches have expanded and diversified over the years. The aim of this talk is to introduce this area of research from a double perspective. First, I describe the various methods that are employed to develop sentiment analysis computational resources, from linguistically-motivated, lexicon-based systems to machine learning algorithms and language models. The tone is introductory and accessible, with a conceptual rather than technical emphasis, although, inevitably, a number of technical aspects need to be discussed. The second perspective focuses on the applications; although sentiment analysis tools were originally developed with a strong industrial bias, applied to marketing research, customer satisfaction or social media analysis, they can also be of great in linguistic inquiry, that is, as a means to explore texts and corpora. Thus, a claim is made that sentiment analysis can be a powerful new arrow in the linguist’s quiver, since these tools can offer insights beyond those offered by traditional corpus analysis tools. I illustrate the claim by describing several actual pieces of linguistic research that use the Lingmotif sentiment analysis tool.