Corpus Linguistics Project

A Rackham Summer Research Grant enabled us to launch a study of student writing that uses corpus linguistics computer-aided tools. Because matriculating students submit evidence-based arguments in electronic form prior to orientation, we have an electronic corpus of over 2 million words of student writing on which we can perform analysis.

Using a subset of the corpus, we have identified the moves that characterize a successful argument, and we will use these findings to develop further markers of effective and less effective writing. Moves we have identified include: (1) reviewing or summarizing the source text (an article students read on which they based their argumentative essay); (2) articulating a stance; and (3) offering evidence in support of that stance. Corpus tools allow us to see, across many student essays, how and when students make these moves and how they differ from the similar rhetorical choices made by advanced and professional academic writers.

One interesting, initial observation we have made concerns the textual complexity that characterizes successful coordination of rhetorical moves. That is, we have found that the more “successful” essays – those essays considered by multiple readers to be “strong” and to fulfill the expectations of the prompt – tend to exhibit more textual complexity, in the form of a balance of what we might call expansive and contractive statements. By this distinction, we mean statements that engage counter-arguments and others’ perspectives, versus statements that assert only the writers’ stance, respectively. In our initial research, it is the students that offer a balance of both kinds of statements that most effectively make a case for their evidence-based arguments.

In the long term, we plan to use these electronic tools to follow a few students of varying abilities to learn more about how they develop as writers.