Discovering language discovering language properties through corpus-based dictionary data analysis

Abstract

To reveal underlying patterns in real language use, linguists have increasingly come to rely on corpus analyses, involving the evaluation of statistical frequencies in generally sizable bodies of natural linguistic data. However, accessing and analyzing large samples of raw language is neither always practical nor even truly necessary, especially in cases pertaining to structural characteristics. In fact, the requisite data can oftentimes be gleaned from a state-of-the-art (i.e., corpus-based) dictionary. Moreover, given the widespread availability of easily searchable electronic dictionaries nowadays, almost any language teacher or learner can use one to answer a number of these types of queries. This paper illustrates this claim with a step-by-step analysis of corpus-based dictionary data for the purpose of formulating the sound-symbol relations in English words with vowels preceding -gh.