Philosophical Discourse in Early New Zealand Newspapers
This project builds on work I carried out as a summer project towards my Masters in Applied Data Science with the UC Arts Digital Lab.
If you open up a history of philosophy in New Zealand, you will typically find that it starts somewhere around the middle of the 20th century, and tells the story of analytic philosophy carried out in academic environments.
I am curious to see what kind of story might be told by applying digital humanities methods to early New Zealand newspaper content. As the image above indicates, there was a lively tradition of public debate and lectures carried out in New Zealand.
My first major hurdle in this project is the fact that philosophical content is a neddle in a haystack within large digital newspaper archives (such as Papers Past). If we were to directly apply topic modelling, or similar methods, to the full newspaper corpus we are unlikely to find any signal from the small proportion of philosophical content.
Figuring out the best way to pick out a corpus of material relevant to a specialised topic in a large digitised newspaper corpus is a general problem. My suggested method for solving it is a kind of ‘iterative bootstrapping’ in which we manually label items and train increasingly targeted classifiers. The code link provides a general outline of the implementation of the method. There is also an online dashboard for generating cooccurrence networks from the resulting corpus of philosophical discourse. A paper on the method is currently under review.
The resulting corpus is, I argue, well suited for investigating disputes over the relative status of religion and the natural sciences within intellectual life and educational policy in early New Zealand. The next step in the project will be to apply the corpus within a piece on these issues as they arose in the work of William Salmond.