Computerised functional analysis of English


On 22nd April 2013, Michael James O'Donnell (Universidad Autónoma de Madrid) conducted the workshop “Systemic-Functional linguistics and corpus annotation” at the Faculty of Philology and Translation in Vigo, invited by the LVTC research group. This workshop guided the participants through the process of analysing a small corpus in terms of Systemic-Functional Linguistics. We used UAM CorpusTool, software for annotation of texts, cross-layer retrieval of segments, and performing statistical studies of the annotated corpus.

Participants were initially shown how to load texts into the software. Then, various layers of analysis were added into the system, one for coding texts as a whole according to their Register (characteristics of the language specific to the situation of production), and others for coding the grammatical characteristics of the text. We explored the annotation of Mood, Theme and Ideation using the tool. Some of this process involved manual annotation (the user selects some text and then assigns features to it). However, the tool does support automatic parsing of text, and this process will be demonstrated. By the end of the workshop, participants were able to perform linguistic studies of their own texts.

Michael James O'Donnell, currently Senior Lecturer at the English Department of the Universidad Autónoma de Madrid, holds a PhD on computational linguistics (University of Sydney). Previously, Michael was consultant and research fellow in a number of research institutions and centres, with funding from the Flemish Government, the Centro per la Ricerca Scientifica e Tecnologica in Trento, the University of Edinburgh, etc. As regards his publications, he has co-edited the volumes Language & Literacy: Functional Approaches (2007, Continuum) and New Directions in Language and Education (2007, Continuum). He has published articles in international journals such as Word, Applied Linguistics or Natural Language Engineering. He was been appointed as keynote speaker in, to cite a few events, the 2nd Finnish Symposium on Functional Linguistics and Multisemiotic Discourse Analysis (Helsinki), the 8th Euro-International Systemic Functional Workshop (Nottingham-Trent), and the 12th Euro-International Systemic Functional Workshop (Glasgow). As a software designer, he has developed packages such as UAM CorpusTool (a corpus annotation tool allowing annotation of multi-text corpora at multiple layers), Systemic Coder (a tool for assigning features to segments of text for statistical analysis) or Workbench for Analysis and Generation (a sentence generation tool, and knowledge representation engine).