The 1970s-the 2000s

Data analysis for Constellate Data for the first three decades of analysis

The 1970s

Public History in its initial phases reflected on itself and saw two things. The first was a potential employment opportunity for graduate students currently in the field of history. The second was an opportunity to correct what many had seen as a failing of traditional history - namely its inability to account for social history and worker's history, of the kind frequently becoming more interesting to the public at large. The 1960s saw a rise in social history through activism and community history, and the larger historical tradition and the academy saw in itself a failing to account for this.

It was quickly becoming apparent that there was a shortage of careers in the history field as a whole. Current graduate students were struggling to find employment, and there was a general distrust and misunderstanding of what historians could bring to a professional environment. In the mid 1970s, programs at UC/Santa Barbara and elsewhere began training graduate students in history to work in the professional field. The goal was to open up jobs in museums, in government spaces, and in the private sector. If students could have more experience in those ways of "doing" history, and if the professional environments in question could see the value in hiring historians for all kinds of work, then the employment crisis would be considerably eased. Historians could find work "doing" history everywhere, and everywhere would see the value of hiring historians to improve their own work.

Alongside this, the field of history was encountering a reckoning of its own making in the wake of the social activism of the 1960s and 70s. History was being done more broadly than ever before, to the wide acclaim of communities, but it wasn't by trained historians. Community history and social history was becoming prolific, and placed its focus on marginalized peoples and groups that were excluded from or underrepresented from traditional historical narratives. In response, many academics wanted to broaden the horizons of traditional history to include these groups, and to produce professional history that focused on all kinds of people and stories.

The data for the first few years of publication in the Public Historian - up until 1980 - reflect the former goal, but not the latter.

There are 300 published articles between JSTOR and Portico before 1980. Using term frequency models, it can be seen that the words "state", "government", "research", "university", "study", "california", and "santa" all appear as some of the most frequently discussed words. This is both the case using the words as a percentage of the articles, and by number of appearances in general. When the generated topic models are examined, the same trend appears. The model generated seven topics, all containing roughly a dozen words each to categorize some of the most frequently used words as connected together. The generated topics and topic lists all support the trend seen in the term frequency model. All the generated topics include a focus on university, policy, government, studies, and work. Words suggesting social history, worker's history, or history from below are not strongly present. Words like "social" and "history" and "oral" are present, but not as frequently nor as strongly as words like "office" and "business" and "training". The focus, as early as the origins of the field as known through its largest published journal, is on training historians to be employable.

This is not entirely unexpected, of course. In many of the earliest articles published within the field, and in many of the later reflection pieces, authors argue for a shift in the historical field to better train historians for employment. It is recognized that there is a need for historians in all sectors of the workplace, and that the private and public sector will benefit from historians. The historian's skill set, as well as the historian's outlook on the world and on important social issues, is valuable, but not seen as employable. The earliest Public History programs, and as a natural result, the earliest articles on the subject, place an emphasis on a desire to see that changed. Historians need better and different training at the graduate level, so that their skill sets and focuses will be more easily accepted by the workforce. It's less a covert operation than a conscious planning strategy to fill what was seen as a need by history grad students and the workforce.

What is interesting is that almost nowhere in the topics for the first few years of publication is there any reference to the goals of social history and history from below. In the topic model lists, when "work" or "worker" appears, it is alongside "office" and "business" and "government". This suggests that the term is being used strictly in reference to the attempt to address the employment problem in history, not in discussion of social history of any aspect. If it had appeared alongside words like "oral" or "social" then perhaps a different narrative could have been construed, but as it stands now the topic model lists collate "worker" with "office" and "federal" and "business". Social history is not the focus, at least not here.

The one outlier in this regard is the term "oral". The term comes up in three out of seven generated topics, and though its surrounding terms still tackle the employment directive of the field, its mere presence suggests at least a cursory focus on the latter goal. Oral history has had a long and complicated relationship with the historical tradition writ large. It is viewed as too fallible to be treated as real history, and conversely its fallibility is viewed as a unique strength within the field. Works such as Portelli's 1991 book The Death of Luigi Trastulli argue in favour of using oral history as "real" history, regardless of how "wrong" it might be. Some public historians argue that getting history crooked, that oral history's fallible nature, it part of what makes it more valuable as history.

Additionally, oral history's complicated relationship with history intertwines with the historical traditions legacy of colonialism. Though outside the bounds of this project to discuss in depth, oral history is frequently the primary method of history for underrepresented historical groups such as indigenous peoples. To view it as not historical, or not historical enough because of its potential fallibility, is in many respects an act of colonial violence. In seeking to course correct history in this manner, public history has naturally over the years gravitated towards oral history as a method of community and social history. The metatextual literature of public history is largely in favour of using oral history to gain a more rounded picture of historical events. It allows historians to see history with a more personal perspective, one that can greatly assist the production of larger historical narratives. Seeing the term and topic "oral" crop up in the corpus as early as the 1970s means that the field has, since its inception, concerned itself with examining oral history. With regards to examining the corpus' data to evaluate its claim of concerning itself with social history, seeing the term and topic "oral" appear with frequency and consistency is extremely promising.

The 1980s

The 1980s saw an incredible uptick in the number of published articles to a total of over 2300. This is not surprising, as the journal was moving into its second decade of publication, and the field was amassing more and more attention. Though this project is relegated to examining the field through its publication in its flagship journal of the Public Historian, it is worth noting that from the field's inception to the modern day it was interspersed with projects across museum spaces, public spaces, and eventually the fledgling digital realm as well.

Compared to the preceding decade, the 2300 collected works contain an overall lesser focus on the world of work. That said, when examining this decades' corpus with tf-idf (term frequency - inverse document frequency) models its references to "government", "university", and "work" are still quite high, and are comparable to the terms "history" and "archive". This shows that though years have passed since public history's first directive to train historians to be gainfully employed, it has still not been surpassed as a focus.

The topic modelling tells a slightly new story, however. Still present in almost every grouping are words like "government", "university", and "California", which harkens to that employment directive that drove the field in the late 1970s. However, in comparison to the 1970s, fewer words in each topic are devoted to this, and some new words are given particular importance. "Research" - as both a word in terms of mere frequency, as well as a topic - shows up frequently. It is present in several of the 10 generated topic lists. It spikes in frequency in 1981 and continues its relevance across the whole decade. "Study" also crops up as a topic and a frequent term.

It begs an interesting question about the usage of the word. Topic modelling and tf-idf models are unable to display context of word usage. This means it is often unclear whether a word that shows up is a noun or verb format, and its relative context is missing. This can impact analysis significantly, and so care has to be payed to all potential versions of context. In the case of "study" and "research", context matters a great deal. In one context, they are being discussed on a meta-intellectual level, and they refer once more to the employment directive of the field as a whole. There are perhaps papers that discuss the need for more studies and research to be produced by the field, and for those studies and that research to be relevant to government and university work to continue to garner employment for historians to better the work force. This narrative is entirely plausible, particularly when taking into account the presence of other employment related terms, and the overall trends of the previous decade.

However, when taken into account with a couple other terms that crop up in the generated topics, there is another context and narrative that significantly changes the meaning. Also present in topic models and tf-idf models are the words "social", "people", "society", and "service". The last word is particularly striking, but all together when taken into conjunction with words like "research" and "study" they suggest an ongoing trend in the field of community oriented research. Recalling the origins of the field, the other directive for the field's emergence was the desire to produce research to formalize 1960s trend of community and social history. Public history was in part meant to help historians enter the work force, and also to produce work akin to a worker's or social history for groups traditionally underserved by the historical narrative and academy. The trends of these words suggest that very likely just such a thing was happening in the Public Historian at this time.

There are two complicating factors that each help to support one narrative over the other. The graph for term frequency that showcases "government", "social", and "society" resembles almost a double helix.

The frequencies of the words twine around each other at roughly the same percentage, indicating that those terms were used in the same number or volume of articles. Though again, this does not indicate context. It is not shown whether the words were used to convey the same meaning, or were related, or whether they were merely present in similar number of articles. That said, the data itself heavily implies a distinct connection. In this regard, the presence of words like "social", "society", "people", and "service" do not imply a growing focus on social history and history from below, but continue to refer to the employment directive of the field. This is not to say that the field itself was not concerned with employment with the goal of producing social history, but rather that the published works were still concerned with history's role in the world with regards to employment as opposed to producing actual works of social history. When this context is understood, there is still not a significant focus on producing works of social history.

In contract, when the term "colonial" is introduced to that same model it could show the beginning of a trend of producing works of social history, instead of merely discussing them with regards to the employment directive.

The term "colonial", though it faces some periods of inactivity, is slowly starting to rise in prevalence over the course of the decade. It is still barely present in comparison to the frequency of the other terms. That said, the term is not present anywhere in the generated topics - it only shows up when the tf-idf models are taken into account. But when the topic models, which contain many ambiguous words as previously discussed, are taken into account with this tf-idf model, a narrative contrasting the employment narrative can be seen. The field was not merely producing works discussing what it would produce, and how best to integrate the skills of historians into the world, but was also producing written research about people and society, including potentially discussing the social history of colonialism.

That said, the generated topic lists still have a distinct focus on employment, and the larger role of historians in the workforce. The topic lists still frequently involve words such as government, research, work, state, federal, national, political, historians, information, and research. The prevalence of these topics cannot be ignored and ought to be taken into serious consideration. There is no presence of a topic filled with words directly related to social history or history from below, and words such as "oral", "social", or "people" require a deeper context that simply is not present in either the topic models or the tf-idf models to fully support the narrative that the field turned in that direction in the 1980s. Instead, though there are markers that such a focus was present, they are few and not strong enough in nature to overcome the other markers that suggest a divided focus of the field that includes dealing with the employment crisis and the identity of historians in the world at large. This is not to suggest that the field had settled on an identity that was devoid of a focus on social history, or one that had leaned fully into simply producing works discussing how best to train historians. Instead, it is clear from the data that the field was still struggling with its identity. It was still grappling with the impetus behind its creation to train historians for different kinds of historical work, whilst also producing work that discussed the nature of the archives, the academy, and even social history such as colonialism. It was a wide and varied field, and the unfocused topics, in which the only common trend is the employment directive, reflect that.

The 1990s

The 1990s is an interesting decade for analysis, both in topic modelling and tf-idf. Once again, seven topic lists were generated in the topic model. Many of the seven contained words like "historian", "university", and "research", linking them in common to decades past. Once again, devoid of context these frequent words and terms could relate to the employment directive, or could be from articles that merely reference the direction of the field as a whole.

The topics do include several terms not seen frequently or at all in the previous decades generated topic lists. Words like "museum" and "social" were there previously, but not frequently, and not with any supporting words to suggest a specific context. Now, they appear alongside words such as "preservation", "cultural", and "historic". These words together suggest a context much more in line with the second of public history's goals: to produce work of a social history nature. Previously, words such as "service" and "social" and "study" were devoid of context and were thus unable to support a narrative of a rise in social or worker's history work. Now, those same words are present, and with whole topics comprising of words like "museum", "preservation", and "cultural", that context is present. Now, those topics can be read alongside a narrative of a growing focus on social history.

This narrative is bolstered when the trends from the tf-idf models are taken into account.

This chart must be read carefully, but when its full context is examined it supports the analysis that in the 1990s more and more articles were being published centred around topics associated with social and cultural history. Though it appears that these terms first gained significant frequency in the 1970s, the data from that decade, through both topic models and tf-idf analysis, does not support that. However, because fewer articles were published in that decade, the percentage of articles that contained those terms may seem higher. In the 1970s and 1980s the frequency of those terms trends downwards and slowly begins to tick up, but in the 1990's those terms, particularly "cultural" and "preservation" start rising in frequency. More articles were published in this decade than in the preceding two combined, yet still the percentage of documents that utilize those terms is rising. This suggests that in the 1990s those terms were of particular importance, and it creates a larger context in which to understand many of the words in the generated topic lists. Words like "research", "study", and "people" before could have strictly referred to articles on the employment directive, but when viewed with this larger context, they could instead be referring to research conducted on people and their cultures.

The other topic of note is the frequency in the use of three terms in the tf-idf model: "colonialism", "oral", and "activism". In previous decades, I have used the term colonialism and oral as a kind of benchmark test. Its presence or absence is an easy way to determine whether there was a consistent focus on social history within the public history corpus. Though many articles can discuss social history or worker's history without using those terms, its presence at all, in any substantive margin is a great indicator that there were academics at that time writing about the kind of social history previously discussed as a potential goal for the field in its original literature. The term "activism" is similarly useful, as the kinds of community history mentioned in the Public Historian's earliest articles refer are also referenced in the context of the activism of the 1960s.

The following graph demonstrates an interesting trend in the data that supports the topic models narrative for an increased focus on social history:

Here, the three aforementioned terms' frequency can be seem throughout the decades of the total Public Historian's corpus. The term "history" is there to modify the results for visualization, as its incredible frequency can demonstrate the actual relevance of the other three terms without them appearing inflated. Additionally, this chart tracks frequency through the number of documents as opposed to in the percentage of all documents, to avoid the potential for false results due to the consistently rising number of published documents over the first three decades. In general, the three terms start to rise little by little, but all see a much more significant increase in the 1990s than in any preceding or proceeding decade. Though "oral" itself is much more present in the literature, this is not unexpected, as it has shown up in the generated topic models for decades past, whereas the other two have not.

This benchmark test is interesting in two specific ways. First, the data of the tf-idf model does show an increase in a focus on social history using these three benchmark terms. The number of articles that use these words in the 1990s is in the hundreds. Though, once again, this is abstracted data devoid of the content of the article, this trend is nevertheless notable. Historians are publishing works at least tangentially, if not completely, focused on social and cultural history. This is a promising supporting note for the narrative espoused in those earliest articles. However, this data only exists if it is searched for. These specific terms are not present in the generated seven topics for the 1990s, and though as stated there are topics connected to social history, the most obvious benchmark terms did not appear in a significant enough fashion in the topic model.

In this regard, the meta nature of topic modelling and tf-idf models comes through. The data can be manipulated in certain ways to tease out specific narratives. I can search for specific terms in the tf-idf model to showcase the potential of a growing trend and narrative, and can look at the connection between specific terms in the topic models as connected to the tf-idf results. But despite my ability to make this connection, I cannot make the topic models generate words and topics that are not present. Oral history is not discussed enough to warrant a significant presence in the topic models, despite its growing presence as evident in the tf-idf model. Activism is discussed more and more frequently over the course of the first three decades of publication in the journal, but it is not important enough to warrant its own topic.

Instead, what does come through significantly in the topics that is unique to the 1990s are words like "preservation" and "museum", which each appear in several topics and thus several contexts. They are connected to words like "development" and "historic", and, in the case of topic 5, one instance of "local". They suggest a focus perhaps on smaller museums and programs that deal in community history. This is not to say that the field all of a sudden began publishing en masse articles about local museums and local history, but slowly in the 1990s, alongside articles clearly still focusing on the university and the potential for history done alongside politics, there were articles about local museums.