Automatically Extracting Typical Syntactic Differences from Corpora

Abstract

We develop an aggregate measure of syntactic difference for automatically find- ing common syntactic differences between collections of text. With the use of this measure, it is possible to mine for differences between, for example, the English of learners and natives, or between related dialects. If formulated in advance, hypotheses can also be tested for statistical significance. It enables us to find not only absence or presence, but also under- and overuse of specific constructs. We have applied our measure to the English of Finnish immigrants in Australia to look for traces of Finnish grammar in their English. The outcomes of this de- tection process were analysed and found to be insightful. A report is included in this article. Besides explaining our method, we also go into the theory behind it, including permutation statistics, and the custom normalizations required for applying these tests to syntactical data. We also explain how to use the software we developed to apply this method to new corpora, and give some suggestions for further research.