Is it the What or the Why

What data am I working with, and why?

As Alan Liu faithfully points out, the analysis produced using topic modelling tools relies on a foundation and a ruleset designed by human hands. What data is analyzed, and on what parameters, affect the results.

It begs the question: what data am I using, and why?

I knew from listening to advice and reading examples that the bigger the data set the better. Ten thousand articles on this topic will yield a more varied result, and thus more likely to be indicative of real substantial academic and philosophical trends.

That does not, however, mean I ought to incorporate as many articles as I can. I had two problems to grapple with: a philosophical problem, and a mechanical one.

Taking the first one first, not all subfields of Public History, interdisciplinary as it is, are going to be useful to me. The field involves ethics and philosophy, museum studies, affect theory, the digital humanities and all its interdisciplinary fields. It can relate to music, to fine arts, to sports and culture, to material and immaterial culture alike. While all of these can fall under the wide umbrella of public history, it is crucial that my understanding of what field I am trying to analyze informs what makes the cut. Just because a journal is on material culture, should I include it for analysis? Will that be useful, or will that confuse my results?

At the core of this philosophical problem rests the idea of what results I am expecting or hoping to get, and how that impacts the data I am obtaining. Though in my initial phase of research I found myself frustrated with Liu's arguments against the objectivity of topic modelling, now more than I ever I agree with him. These small decisions have real weight to them, and can seriously influence the data I am researching.

My understanding of public history comes largely from David Dean's work A Companion to Public History. I took his class, and listened and spoke with him for several months on what Public History means and what it could mean. How does memory and trauma relate to it? What about colonialism? How are protests any different a form of public history than museums? Is there even a journal of articles relating to protests then, for me to incorporate? Are those journals of affect theory, or journals of American politics? The frightening thing is that this strain of thought can be applied to each and every subfield that relates, even tangentially, to public history. It's a big field, it can be a highly theoretical field that, at least from my conversations with David Dean, can and should always be changing.

I personally view it as a field that emerged out of the more general historical tradition as a way to create history with the public acting as audience and co-author. It was about the ways the public engaged with the past, and the way the past is understood by the public. Of particular concern, to my understanding of it, is the connection between history and memory. Is it faulty and imperfect, and how much importance is that imperfection? But again, the field is interdisciplinary, with as many subsidiaries as you can think of.

That said, there is a methodological problem here, that can answer some of my philosophical problems. After all, my belief in what the field of public history is is not just a philosophical query in which I spend a few thousand words to address it, but one in which the parameters I set for this project answer it for me.

I can download articles either by topic and keyword searches, or by journal. The Journal of Public History - a shoe in for this project if there ever was one - has 10,000 catalogued articles available for download. Keyword searches for public history with a date frame of 1970-2020 number somewhere north of 2.5 million. A little beyond the ability of the Constellate Beta to download, and probably not a useful dataset to download.

I can specify by topic - JSTOR's catalogue of keyword "public history" + 1970-2020 + history subtopic has only a little over 20,000. A manageable amount, but that is just one topic, subcategorized to high heaven. But I can download by journal, and there's a perfectly ordered and organized Journal of Public History right there.

It's worth returning the question of data sets a little later. A heavy philosophical problem needs to be addressed there, and the resolution of it will define in part the utility of my results.