The identification of argument schemes in natural language text is a time consuming process. However, through the use of more modern machine learning techniques trained with previously annotated argument maps created in Ova+ (a browser-based tool by ARG-tech) a classifier can be built to automatically identify these schemes.
This tool takes plain text and firstly separates argumentative from non-argumentative speech. The argumentative segments are then classified with a degree of certainty for which of Douglas Walton’s argumentation schemes it most closely represents. The classifier itself was given an labelled training corpus consisting of arguments made during select committee oral questions. This is mostly suited for analysis of semi-formal or political text and is another step in creation of tools for argumentation research.