Where to start?
A generic overview of the four tasks you need to carry out for discourse analysis using corpora and corpus linguistics tools is shown in the image to the left. While the overall process is indeed that simple, each of these tasks comes with quite a substantial load of subtasks, some of which can only be completed after some of the other steps have yielded necessary information.
Ponder, for example, the implications of the task “build and pre-process corpus”: For corpus building, you need to decide which texts are suitable for your project. Not only does this imply some sort of research question your texts should make possible to answer, it also implies that you must have a means to determine whether any given text contributes to answering it or not. In other words: you need to analyse texts to be able to deem them “suitable”. And that means you have to carry out the task “analyse texts” while you are actually still in the process of finding suitable texts.
Obviously, you need to loop through those steps to make sure that anything you do is contributing to the overall goal you are trying to achieve: completing a research project.
The keyword I would like to introduce here is “pilot study”. A pilot study should be a micro version of your study to make sure that there will be no unpleasant surprises in the middle of your project.
Pilot Study: Taking it one non-linear step at a time
Find a research question: Subtasks and actions
Discourse analysts are interested in finding out and showing how language use shapes our way of thinking and, by extension, our perception of the world (and, some may argue, our world itself).
To find a research question within this paradigm you need to iterate through the following steps:
- Check salient discourses [→ read and take notes]
- Primary sources: What is being debated right now?
- Secondary sources (especially published research): Which discourses or which perspectives of discourses have not yet been analysed?
- Check theories and models used in discourse analysis and corpus linguistics [→ read and take notes]
- Which models and theories are best suited for what kind of analysis?
- Which ones could you apply, challenge and/or develop within your research project?
- Which ones could you apply, challenge and/or develop within your time frame?
- Write down your thoughts [→ write in full sentences, take notes on what to research and read next]
- Describe what you plan to do
- Write parts of your thesis (resp. try to) and make note of where you got stuck. Anything you would like to write about but cannot you need to research and read up on.
- Writing things down helps tremendously to find out what is still missing. If you cannot express what you are trying to do, you probably don’t know yourself yet.
Let’s assume that you are interested in the discourse on climate change. With this research interest in mind, your next step should be to get an idea about the communicative events that deal with this subject (i. e.: find suitable texts). You can (and should) approach this from two perspectives: 1. Take all your knowledge about communication into account. 2. Do not presume prematurely that your knowledge about communication will lead you to all communicative events that are necessary to carry out the study. For the discourse on climate change, your knowledge about communication tells you for example that this subject will be covered in media texts.
[to be continued]