Non-Standard Syntax Annotation

3 Credits · Level: Graduate, Undergraduate

Annotated corpora and corpus methods are increasingly being used as a source of empirical evidence in syntactic research - due in part to the availability of larger corpora and more detailed analysis possibilities. At the same time, the focus in syntactic corpora has been very strongly on canonical English, particularly newspaper language, which is used as a basis for automatic Natural Language Processing. This course aims to extend the borders of standard syntax analysis in corpora to non-standard varieties of English. We will learn two different frameworks for carrying out syntactic analysis of corpus data, also referred to as 'tree-banking'. The course will accordingly be divided into two parts, devoted to analyzing constituent trees based on the Penn Treebank guidelines and dependency trees based on the Stanford Typed Dependency guidelines respectively. In both cases, we will begin by learning the guidelines for standard English, and then apply and adapt them to a non-standard variety of data, such as dialectal English, World Englishes, non-native learner English, historical varieties, language on the Internet or disfluent spoken data.