Conclusion

This essay, and this associated website,, set out to trace over time self-identified trends in the early days of the field of public history to determine whether the two stated goals of encouraging employment and generating local and community oriented history were present through the field’s history. To do so required a combination of close and distant reading, which was achieved by employing digital humanities techniques and technologies, specifically topic modelling, term frequency – inverse document frequency, and raw and relative term frequency models. By utilizing these methods, I was able to take 10,000 articles – the entire digitally archived corpus of the Public Historian journal – and use Constellate’s Beta digital workbench and Jupityr Notebooks to produce topic modelling and tf-idf and related analysis. This data was guided by and supplemented with my own historiographical research into public history.

The early works into the field of public history suggest a clear focus on what I have referred to in this paper as the employment directive. Essentially, historians saw empty spaces in the work force where historians could and should be, and where they could positively impact all kinds of work with the ideals and methods of historians. But historians were not being hired. The fix was to create graduate programs in North America in the new field of public history. Historians would be trained to apply their skills and methodologies to different kinds of work, and would have work placement programs meant to ingratiate themselves into the workplace.

The data I analyzed from the first few decades of the journal’s publication reflect this clearly as a focus. Up until the 2000-2010 corpus, the first three decades had topic lists generated that overwhelmingly referenced the employment directive. There were different emphases, of course. Some topic lists had words in such a collection that I would label them “research work”, as they contained references to researching, writing, the university, and studying. Some were clearly what I would label “government work” or “private sector work”, and contained references to all levels of government and to offices. The tf-idf models generated out of these corpus’ of the first three decades support this, or at least do not contradict it in any meaningful way. It is clear that the focus of the several thousand articles published at around this time were solidly focused on producing work analyzing the field, and what and how it could offer services to other sectors of society. This was a self-declared goal in the field’s early literature, all the data and analysis confirmed it.

What was interesting was that even in the first few years of publication, the topic models identified the word “oral” as cropping up both significantly in terms of itself and in its relation to the other words in that topic list. This is a clear reference to the other, self-stated goal of public history. Early on its literature, the field identified itself as coming out of the 1960s social revolutions, in which various community groups were functioning as historians. The field saw this as another space where historians could lend skills and methodologies to assist in this practice, and to bring to light historical narratives often underserved. This goal is not significantly represented in the first three decades of the field’s publication according to the topic models. In the first three decades, there was no topic that I would label as “community history,”

“cultural studies,” or “worker’s history.” Instead, there were small instances, like the term “oral” in the first decade, or “service” in the next, which begin to suggest an additional focus within the field. This was where I turned to tf-idf and raw and relative data frequency models to analyze the corpus’ with a new lens to provide the context that my topic models lacked.

The topic models I generated lacked context; they functioned as distant reading and provided a large overview of the broadest strokes of the field for my analysis. They could not assign deeper semiotic meaning to the topics, and as argued the topics themselves functioned more like discourses. But by using tf-idf and targeted term searches with raw and relative term frequency models, I can supplement this digital tool to provide a more complex and rounded framework through which to analyze public history. Through all the decades, terms that to me have become hallmarks of modern public history could be searched to see their frequency within the corpus’. Terms such as “oral” as in oral history, “colonial” and “colonialism”, and even “community” are connected intrinsically to this other self-stated goal of the field. Their frequency alone would not show that the field was valuing and producing this kind of work, as the context behind the word’s usage is missing, but it could suggest an overall trend for me to track.

This was exactly what it did. Early on, tracking these terms yielded a non-result. They were not present at all in the generated topic models, and not present in significant quantities in the term frequency models. To me, this showed that the field was not producing significant quantities of work centred around the idea of community history. Though it was present as a topic of discussion in the meta-content of the field, it was not present in the data of the content of the field. Slowly this began to trend up, but it was not really until the 2000s that it took off.

Starting in the 2000s, the topics generated by the topic model flipped the script. Now, topics about work and studying and the university were in the vast minority, and topics about museum studies, cultural studies, and the people or public were in the majority. Still, some words like “colonialism” and “oral” were not present in the topic models. This to me, meant that I needed deeper context to understand what exactly the new direction of the field was. By supplementing my analysis with tf-idf and raw and relative frequency models, I was able to target those terms and others that appeared in the topic models to determine their contextual relationship, if there was one. What I found was that there was a relationship. The hallmark terms of the community history directive were on the rise, with easily double and in some places triple the frequency they had in other decades. They were connected to a rise in frequency for words like “exhibit” and “archive” and “community”. Additionally, more works than ever were discussing the importance of “people” and the “public”. Though the latter is not a surprise considering the nomenclature of the field, what was encouraging was that the term “publics” was on the rise as well. The field was not only discussing community history and different ways to conduct it, but was self-reflecting on its potentially duty to represent the public as heterogenous entity. The new focus as of the 2000s seemed entirely concerned with using museum studies as a vehicle for community and local history.

This is not a surprising result when taken into account with my research question. Both of the goals stated by the field in its earliest literature were present throughout the journal’s publication. Early on in the field’s existence, the literature had an overwhelming focus on the employment directive. It produced work discussing the fledgling graduate programs, and the role of history in all kinds of workplaces. But in the 2000s there was a shift to producing work that focused in on community history and, overwhelmingly, museum studies. Grele early on argued that the field could be well suited to produce work of this nature, but that it simply was not set up to do so. The data proved him right, and eventually it went on to prove him wrong.

The data and my analysis of it is not surprising, but it is still significant. This broad of a metatextual perspective on public history has not yet been done, and significant strides in the analysis still need to be made. What I have accomplished here is to take a macroscopic perspective on the field to take its claims seriously, and I found the raw data to back them up through interacting with topic models and term frequency models, including confirming the timeline on when exactly that shift in focus occurred. The field is no longer new, having now been producing works in its flagship journal for half a century, and as several recent articles suggest, it is time to start reflexively examining the field to see where it has come from and to start pushing it forward with that in mind. But my focus on the data was as narrow as the field’s self-stated focus. I was bound by the goals the field set out for itself and to examine those trends specifically. But there are other trends present, and the work I have done can be replicated and changed to analyze it to answer different research questions. The data can also be viewed through different lenses for analysis by changing the margins of what counts as a public history publication, or the parameters of the topic models can be changed to produce different results. The macroscope that I used to analyze this corpus for this result can be changed, slightly or dramatically, to analyze the same data again to see what new results might occur. A focus on the history of oral history within the field could be analyzed, or a focus on the way public history has engaged with museum studies, or worker’s history could be produced. Looking at public history in the way I did functions like a microscope. Multiple lenses of different grades are used, and when used together they magnify the subject for close examination. By using close and distant reading through both topic modelling and term frequency models, I have produced a complete view of the field of public history in terms of its early goals, and these same methods can be applied to the field to answer any research question, no matter how big or small.