This talk is basically a demonstration of how quantitative methods of different degrees of sophistication can inform linguistic research on various levels of linguistic analysis. I will report on three case studies. First, I will show how very simple statistics can be used to explore aspects of Spanish Internet Orthography, specifically how standard spellings are changed in online forums and comments and how even speakers’ typing is influenced by semantic and articulatory characteristics of what they are typing. Second, I will address a frequent question in historical data, namely how to study morphological change given the inherent noisiness and multidimensional nature of the data using exploratory as well as hypothesis-testing statistics. Finally, I will discuss a fairly new method designed to facilitate the exploration of how speakers of a certain kind (e.g., non-native speakers or indigenized variety speakers) differ from a ‘standard/reference’ group of speakers even when human annotators of, say, learner data are not available.
